Sapphire 11293-01-20G Radeon Pulse RX 5700 Xt 8GB GDDR6 HDMI Graphics Card review
Sapphire 11293-01-20G Radeon Pulse RX 5700 Xt 8GB GDDR6 HDMI Graphics Card review
pros
Slightly faster than AMD’s own Radeon RX 5700 XT, and roughly 10% faster than GeForce RTX 2060 Super
Significantly cooler-running than the reference Radeon RX 5700 XT
Dual-X thermal solution operates quietly
Dual BIOS
Trixx software is clean and functionalExcellent gaming performance
Cooler and quieter than reference designs
Trixx Boost software uses upscaling for great FPS gains
Dual-BIOS for performance or acoustic focus
PCIe 4.0 compatible
cons
Consumes more power than the reference Radeon RX 5700 XT, which was already higher than GeForce RTX 2060 Super
Large cooler slightly overhangs third expansion slot, making placement an issue
MSRP means you’ll pay more for a Pulse Radeon RX 5700 XT compared to the least-expensive 5700 XTs and GeForce RTX 2060 SupersPlastic shroud
No dedicated ray tracing hardware
Trixx Boost feature not available at launch
SPECS OVERVIEW
Model
11293-01-20G
100416P8GL
Model
11293-01-20G
100416P8GL
Series AMD Radeon RX 5700 Series
GPU Interface PCIe x16
GPU Chipset Radeon RX 5700 XT
GPU Memory Size 8GB
GPU Memory Type GDDR6
GPU Memory Interface 256-Bit
GPU Clock Rate 1670MHz
GPU Boost Clock Rate 1925MHz
DirectX DirectX 12
OpenGL OpenGL 4.5
TDP 241W
Length 254mm
Cooling Type 2 Fans
Power Connectors 1 PCIe 8-pin + 1 PCIe 6-pin
SLI Support
CrossFire Support
DVI-D (Dual Link) 0
DVI-I (Dual Link) 0
DisplayPort 3
HDMI 1
G-SYNC Support
FreeSync Support
the rx 5700 series launched were related to the cooler and the drivers the cooler had serious thermal and noise challenges expected of a blower design and it ended up running unnecessarily hot and loud for a GPU that was otherwise competitive with Nvidia the driver issue has only had partial resolution but overall partner model cards have been expected to solve most of navvies launch issues those relating to thermals and noise and today we're looking at sapphires rx which you saw 100 XT poles card to see if that thermal and noise scenario has
been resolved before that this video is brought to you by Nord VPN Nord VPN is what we use internally as our Virtual Private Network solution we use Nord VPN for international travel to bypass firewalls and at home for data privacy especially from Snoopy and Internet Service Providers Nord VPN helps secure and protect your network privacy and is currently offering 75% off of a three-year plan with one month free claim the offer at Nord VPN comm slash
gamers Nexus and use code gamers Nexus at checkout or click the link in the description below sapphires rx with you 700 XT pulse is more than the reference model that be this card so according to the Cree release pricing we got at more you end up with a cooler that's got a dual fan axial approach it has two 100 millimeter fans so they're large enough that they can spin slower to move more air than a smaller fan and therefore remain quieter and then it's
also got a fairly large aluminum heatsink on a copper cold plate as expected there's a base plate and then the fins are oriented horizontally so about half of your exhaust will go out of the case the other half will remain in the case to be scooped up probably by the CPU fans as opposed to vertically oriented which a lot of coolers are where they go this way and exhaust either up and out towards the glass or down into the motherboard so the sapphire pulse looks good
everything about it looks like a real cooler and a real proper video card design today we're looking at that one we have benchmarks that are really heavily focused on thermal and noise so we've got 40 DB a noise normalized testing versus this card the stock version and then we also have frequency testing so we're looking at the a stable frequency under a heavy workload over a period of time about 30 minutes and that allows us to see does the frequency decay as a result of thermals separately this card the pulse has to be bios's so it's got a switch that's a big
upgrade we liked dual V bios because it gives you an out if you end up breaking something and it also means that there's some modding capabilities that are opened up because again now you have a safety net these second b bios is a slower one the default v bios is towards the front of the card and that's going to be a bit more aggressive fan curve as we'll talk about in the testing and also therefore runs a higher frequency because it can run cooler and the other v bios the non-standard one is quieter it sticks to a much lower fan RPM but also will run into more thermal
issues if you're running in the hotter case so that's the basics of it it's also supposed to be priced instead of the MSRP of the reference card but at ten dollars more expensive for what looks to be a much better cooler and as you'll see in the testing is a much better cooler it's really not that big of a price jump let's get straight into the numbers so we can just talk about how much the cooler really mattered for
this card because that was the big hold point we had on the 5700 XC reference design the reference design is just straight not good we would not recommend it and so sapphire and the other vendors coming out today should start to change the story to be focused more on actual GPU performance and less focused on thermals and noise will go through thermals and noise frequency testing first and then get through the games as always and then we'll have a separate
teardown over the card as well if you'd like to see the cooler design that'll be in a review uploaded on this site as well so subscribe for that let's get into the thermal and noise testing our first chart shows thermals at steady-state after a 30 minute power virus workload measured in a constant ambient temperature of 21 degrees Celsius and monitored every second of the test the Sapphire rx 5700 XT poles clearly solved the reference 5700 x DS biggest
problem outside of the launch driver functionality and that's the problem of thermals with both cards locked to a fixed 40 DBA our noise normalized plots show the thermals for the Sapphire poles GPU edge temperature at 67 degrees Celsius which is a massive twenty seven point one degrees Celsius cooler than reference 5700 XD at the same noise level there noise normalized here they sound the same that's a crazy drop if you ran the 5700 XT under auto settings instead at least with the launch drivers you'd have found a GPU edge temperature of 79 degrees Celsius in exchange for an 11 DBA increase in noise levels decibels are logarithmic and human
perception is hard to objectively measure but in our consultation with sound experts roughly every 10 DBA increased results in a 2x perceived increase in noise to the human ear that's not the same as the increase in acoustic power it's the perceived increase in loudness to a human very different things so just to be extremely clear on that not talking acoustic power GPU Junction temperature which is the governing temperature for throttling and frequency plotted at 110 degrees for the reference card at 40 DBA that's t.j.maxx we're throttling anything
beyond this we can't actually measure because the clocks are now dropping so it would be hotter if we were allowed to run the frequency at the same level the Sapphire card hits 83 degrees giving more Headroom for lower noise levels if you wanted them at any given frequency and temperature that's also a higher frequency for reference even full-auto on the blower cooler the original resulted in 95 degrees for launch memory modules are well under control on the Sapphire card at 40 dB a plot in 82 degrees the reference 5700 X T's 98 degree
result is in questionable territory where G DDR 6 really shouldn't be operating finally MOSFET temperatures around 68 on the sapphire pulse at 40 DB a or 78 on the reference card at an equivalent noise level if you blot the noise level to annoying volumes it's about the same but noise normalizing really matters here these vrm MOSFET numbers are fine for both cards ultimately but Sapphire is doing well at 40 DBA b-roll of the Sapphire card makes it pretty
obvious why it's doing so much better the two fans are big and there's two of them and they're axial so those are all good things so far the Sapphire card is massively improved over referenced for thermals which was our biggest complaint the first one with the reference card a few people out there tried to argue that the reference blower wasn't actually loud or hot but this is why that was wrong it was both objectively hot and objectively loud particularly in the comparative against partner models there's no point in defending a bad cooler design and
Sapphire gives us the numbers to definitively show that there is a better way to make a car at about more than the reference blower it's not even that much more expensive the next chart shows the frequency response of the sapphire rx 5700 XT for each V bios it's got two switches but also compares to the reference 5700 XT AMD card starting with the pulses default V bios the right switch position the Sapphire card holds a range of about 18 80 megahertz the 19 10
megahertz bouncing around hard as a result of aim these various power voltage and thermal thresholds and because it's being pushed pretty high already for comparison the reference 5700 XT by AMD held a range of about 1860 megahertz to 1906 megahertz so they're pretty close to each other with the average closer to 1880 maybe 1885 the reference card bounces around less dramatically but also maintains a lower overall average GPU core clock and for reference the
GPU core clock will change depending on the workload so this won't be the same in every single game or application but we're using the same application tested the same way for each of these cards so they aren't directly comparable finally for the left v bios switch on the poles frequency starts around 1830 to 1850 megahertz but Falls with the rise and thermals to a range of 1815 to 1840 megahertz averaging around 1830 this directly correlates with performance of course so you'd run a lower fps in exchange for the lower frequency but also a lower frequency it means
the card isn't working as hard but left to be BIOS is also less sporadic and more stable in frequency but it's lower performing overall for how this translates to fan speed here's the next chart with just the pulse b bios switch positions the reference card isn't shown since it's a different fan so it direct rpm the RPM comparison isn't linear and doesn't make sense fan speed will be chosen along an RPM temperature curve so HP BIOS has a temperature target then it adjusts the fan to meet the target the pulse start with the default bios tries to maintain
the GPU temperature target of about 71 degrees Celsius edge temperature not Junction temperature during load plotted here the left be BIOS allows more silence and forces a temperature limit about 74 degrees Celsius but it also down clocks so it'd be running hotter here much hotter if maintaining the same 1900 megahertz clock as the first default V bios as for how fan speed is affected the default v bios pegs the pulse fan RPM to about 1570 or just under 40 DBA at 20 inches in a room with a noise floor of 26 DB the left v bios the quieter one pegs
RPM to about 1000 after an initial burst period this heavily controls noise level and brings down the card to about 33 to 34 DBA on an open-air bench at 20 inches and your average computer case you'd never hear this but it also run at least 5 to 10 degrees hotter and an average computer case to be fair we wouldn't recommend this v bios which for any one putting the card under routine heavy loads but it'd be fine for a user with less heavy usage or a case with heavy airflow focus or direct to GPU focus like you know 11 dynamic or something like
that the next question is how the two different view bios has impact PRM and memory thermals we do a lot of stress testing for thermals with fir mark but for this we're just using a constant 3dmark workload that generates something more similar to a game not to be annoying but again remember that this is with lower frequency on the Left V BIOS so the numbers aren't perfectly comparable but this gives us an idea with the default V bios the GPU edge
temperature averages to around 71 degrees in 3d mark or nearing 74 degrees with the other V bios G V Junction temperature the hottest single spot on the GPU is the same in each test roughly 87 degrees this is with ambience accounted for and controlled 221 degrees has a reminder the memory modules show a big difference memory on this card benefits a lot from the fan so droppings 1000 rpm increases the already hot GTD are six modules up to about 92 degrees Celsius in a hot case that'd be entering dangerous territory for memory so we'd again
advise against using this in the hot cases for heavy loads the VRI mosfet temperatures also climb about eight degrees but we're still well within a reasonable operating range Sniper Elite 4 will start us off for games overclocking worked out a bit better with a sapphire card than with our reference 5700 XT but that's because our memory was better on the Sapphire card the actual and the overclocking process is still either broken or just will never have that much Headroom to begin with this is luck of the draw in the memory because the memory on our
5709 xt was also better than our xt but either way we'll take the games full stock the Sapphire 5700 XT pulse ends up at 71 FPS average with the default P bios 60 9.5 FPS average was achieved with the quieter V bios but note that this would be worse in a hotter case and both of these are setting our frame rate as within run to run variants of the 5700 XT reference cards 71 FPS average there is no improvement at stock when it comes to raw frame rate but again
there's a bigger picture of noise and thermals with lower thermals that it give a noise level you can get the same experience for a significantly quieter and that does matter further overclocking the memory worked better and sniper than it did in some of our other games where we were able to hit nine hundred fifty megahertz instead of just nine hundred that we did on the original reference card although in a lot of the other games we tested we did have to
drop down to 940 or a 935 to get stability and everything the 5700 XT pulse at 78 FPS average when overclocked with a low scaling similarly and demonstrating that the overclock is stable in this game 78 FPS average gives us a gain of almost 10% over the stocks sapphires 5700 XT poles or the reference card but this is the last title where we'll see it gain that big from overclocking at 4 and 2018 is next giving us a DirectX 11 classic look rather than the previous dx12 an async compute title at 1080p the RX 5700 xt pulse maintained 151 FPS average which
is within runs run variance of the rx to the 700 XT reference card by AMD lows are also within the wider variants so there's functionally no difference in gaming performance when ignoring thermal acoustic and general quality of life improvements for the Sapphire card toggling into the quieter V bios brought us down to 149 FPS average and overclocking brought us up to 156 FPS average on these sapphire 57 or XT with a 9 50 megahertz memory clock on this title at
1080p although we dropped it at 14:40 and forty a and that resulted in a gain of 3.6 percent over the stock 5700 XT pulse at 1440p the 5700 axial managed to one 15.7 FPS average which is well within the variance of the reference one fourteen point nine FPS average lows were also within variants the quieter v bios Bryn's performs down to one thirteen point six establishing that the default position runs about 1.8 percent faster than the quieter position that said even the default v bios is significantly quieter and cooler than the 5700 AXI reference card so you
don't really even need the quiet one to get an improvement over that overclocking allowed the poles to increase to 121 FPS average a four point seven percent gain entirely thanks to that memory bump at this point we're not sure if the core frequency knob is limited and usefulness broken if na'vi is limited or if the drivers need improvement but we want to see more improvement out of overclocking than we're getting it seems like for the most part this is going to come down to how good is the memory you got on your model and unless any manufacturers
guarantee in a memory clock there's not much you can do about that na'vi really needs that faster memory and seems to follow the trend of AMD cards being memory bound more than anything comparatively in RTX 2070 super is encroached upon by the overclocked 5700 xt cards and the 5700 xt card also performs outperforms the 2060 super which is at 104 FPS average in the stock position at 4k the 5700 xt pulse established that it's fully capable of 4k
gaming with the right gains f1 is one of the easier ones to run although lows begin suffering as we pump the resolution the reference card 68.9 FPS average and the poles at 69 point 4 FPS average are with invariance once again but overclocking bounce the 5700 XD poles it's a 74 FPS average we're ahead of the 2016 Subaru an overclock and stock the stock behind the 2070 super without any tuning the 2070 Super Stock maintains a lead of about 6% over the pulse when both are set to stock settings shadow of the Tomb Raider with DirectX 12 is next 1080p has the
rx100 XT at 124 FPS average for the reference model with the pulse also at about 124 to 125 there's no straight benefit once again other than other quality of life improvements that we discussed earlier it's not like just buying a partner model will immediately improve fps it does depend on the card of course some of them might pre overclock higher than others but ultimately you're looking at a scenario like this where it's about the same out of the box except for the things that we thought really mattered like thermals and noise we dropped the
overclocking numbers at this point - because Shadow the Tomb Raider with any overclock on our pulse at all would trigger a hard shutdown just a black screen it's the same behavior as OCP would trigger which is odd because there wasn't any OCP issues so this is an AMD driver issue and even dropping down to 9 20 megahertz memory caused the same hard crash so we end up giving up on overclocking numbers from this point on and just focused on the basics at the end of the day if all we're doing is overclocking the memory a bit it's kind of the same as
our if it's 200 XT overclocked so there's no point in doing that again what the reference model that is it seems AMD still has a ton to fix and it's broken overclocking behavior for the Navi series cards but at least Sapphire picked up the slack on the cooler at 1440p the rst 700 xt pulse runs at 85 FPS average which is about the same as the 84.5 fps of the reference card the left v bios drops to 83.4 FPS average allowing the default v bios a lead of 2% on the poles card
nothing else here is really worth talking about since it's the same numbers as the reference review aside from noise and thermals there's a clear trend here so no point in showing 4k numbers go check out reference 5700 xt review from launch day if you want the rest of the game benchmark numbers for shadow of the Tomb Raider let's move on strange Brigade with DirectX 12 is up next as a reminder Andy's been doing better in DX 12 the some cards that strange Brigade then with Vulcan at 1440p the 5700 XT Paul stocked the bios frames at 120 4.5 FPS average within variants of the 120 2.8 FPS average of the reference card the left v bios
drops down to 122 FPS average just refined of reference we're going to stop at this point we have numbers for Strange Brigade at 4k with Vulcan we have numbers for hitman 2 and D X 12 and dx11 at 3 resolutions each we have Far Cry 5 and 3 resolutions GTA 5 and 3 resolutions but none of it matters at the end of the day we're all FPS is equal between reference and the pulse so you can check our original review for the rest of those numbers this is all about thermals and noise all that matters for this review other than maybe some follow-ups on
overclocking the game results clearly there's no real difference we have all the other data as stated but if you want the rest of the FPS data you should just check our initial 5700 XT reference card review because the FPS numbers are the same so we're looking for FPS comparison - and video or other AMD cards check that review but for this one it's the same performance as reference so what's really improved is everything else which is important the noise levels you can now run the card at 40 DBA and not incinerate it it's almost 30 degrees cooler with 40 DB a noise normalized the test setup and even if he ran the original one without
that constraint of noise normalizing and a lower fan rpm you'd still be running at like 10 DB a higher than this thing can for the same frequency actually this gets a higher frequency technically even with that noise reduction so Sapphire has done a good job with its pulse for the 5700 XC the cooler is good thermals and noise levels are drastically improved memory temperatures are also dropped to the 80s from nearly 100 degrees Celsius at 40 DBA originally which is a massive step up so if you're buying an RX 5700 XT for sure if you if you've already
decided you're buying one of these cards and it's just a matter of what card do you buy we have no issues recommending the sapphire pulse it's done well it objectively is just massively improved over the original the reference model and we have no issue recommending sapphires version of this cooler we haven't tested the other cooler so of course it's possible unlikely that there are better ones on the market but for what we know today at more than MSRP this one is not only majorly better than the reference design it's probably in the good enough
territory where anything that's improved over this cooler might not be meaningfully so so that's kind of where we stand on it and again we might look at more models later if there's interest in a specific one let us know below and maybe we'll buy one of those now the biggest downside here is still it's on AMD this is not a sapphire issue at this point it's a name the issue and it's the drivers the drivers have been improved so they've updated a few things like that issue with the coloration that we complained about in the original review where the out of box experience for
this thing when we first got it was you install the drivers on any one of I think we tested it on three or four different windows installs and systems and every time without fail after driver install the screen just splashed a really bright green or pink or something and required to restart and it was solvable by restarting it but that's a terrible out of box experience andy has fixed that that's fantastic that's really good news that's an issue that never should have been
there to begin with so it's it's kind of inappropriate to give praise for fixing it good job AMD you fixed an issue that shouldn't have been there and that actually is a pretty big improvement in our books now unfortunately there's still other issues drivers are in a state where they're sort of good enough we didn't encounter any issues when running stock from just one test to the next we were able to run the whole bench suite without any random crashes things like that but there are still issues and most of them are quality of life things like trying to actually change
anything in software with any software not just wot man afterburner as well as this is an AMD api issue but a couple of the issues as quickly as we can recap them I did one take of this already almost 15 minutes long but one of them setting a fan speed does not apply under idle so if you want the fans to spend at forty percent at all times it won't actually start spinning until there's GPU load now we view this as an unexpected behavior for the software if the user sets a fan speed override by nature of being and override the fan should spin at that speed at all times
period so for fan noise test and we actually had to run 3d mark in the background just to change the fan speeds you can't change it on desktop it won't do anything and normally we would do it idle from desktop also there's an issue with reporting sometimes the software report frequencies that are impossible to hit like in the tens of thousands there's also a small bug that's not really too important but at zero rpm mode the fans you'll see them not spinning but softer
reports it as a 4% fan speed like 200 rpm which is obviously not correct setting a higher fans feed is an issue as well you can't bring it back down really so this has been around since Vega frontier edition it is incredibly frustrating that Andy hasn't fixed this yet we have privately complain about it and then we publicly complained about it's still there if you set a fan speed of let's say 70% 80% 90% doesn't matter and you try to bring it down you basically have to
restart to change the fan speed or weight an extremely long period of time we walked away once for half an hour and it reset itself but this has been a bug for a long time it exists an afterburner existed want man it's an API issue or something we want to see it fixed and then also the new driver has black screen hard crash and lock issues that resemble OCP but we don't think it's OCP because that doesn't make any sense this is just one overclocking the Sapphire card this is not a sapphire issue this is an AMD driver issue but basically that problem we run into is when applying even a light overclock of let's say 10 megahertz to memory even
though we know that stable it'll survive eight hours burnin with like a 40 megahertz all increase in 940 megahertz even though we know 10 megahertz extra is stable in some games seemingly randomly we're not quite sure it'll just snap it just shuts down and you have to restart so it's a completely graceless really hard crash and that's bad Andy has this issue where with overclocking not sure if it's just a bug in general or it's some instability that's encountered but and these failures with overclocking when they happen are graceless and so the software will
bug out and be unusable and you'll have to restart or it just shuts down on its own you have to turn it back on or it restarts hard boot so this is an issue where Nvidia does a good job with graceful failures where unless you're under ln2 or pushing extreme clocks for the most part when you're testing stability a clock fails you just restart the application you don't need to restart the whole computer this is something we think AMD can improve upon and we would really like to see it but they seem to be behind still on drivers and on software in some ways
where other parts of their software like the video capture stuff the relive suite that's really good it's just that the poor driver suite needs further improvements overall though we think that they're usable you can play the games some people most probably want to ever notice these issues and then those of you who do we just want you to be aware of them even though you know you can work with them you can live with them it's just kind of annoying quality of life
stuff now let's get back to the roots of this this card fires pulse 50 some 100 XD is very good we have absolutely no issue recommending the pulse cooler sapphire has done a good job with this cooler and it performs well for memory performs well temperatures that is performs well for the MOSFET temperatures performs very well for GPU edge and Junction temperatures so sapphire a job well done if you're buying at 5700 XT at $10 more than MSRP absolutely no
issues recommending this we are happy to say this one's passed our tests now as far as the GPU the thing that AMD makes the 57 rx T the GPU competitively with Nvidia is in a very strong position now the 2060 super is routinely beaten by it the 27 t super tends to hold the lead on average over the 5700 XT few games so that's not true but on average is a bit ahead it's just that that lead is getting small enough now where the price difference doesn't necessarily justify
it so AMD is competing on price as it always does it's in a good spot in terms of price versus Nvidia and for those reasons we can say that the 5700 XC with a partner model not a reference one is a card that we are fine with you purchasing and using for gaming workloads we just ask that you be very aware of all these issues that primarily affect things like again the Fraser is quality of life meaning that it's not we're not running into catastrophic issues that prevent you
from using the device we're just running into issues that are kind of annoying but you can work around them or just ignore them like with overclocking we're eventually like you know what it's 4% if we get it working screw it I'm walking away from this it's not worth any more so you may end up in that boat where it's like you know what is it worth fighting software every day
for 4% maybe no not really so it's it's not like you're losing a ton there but it's something we'd like to see fixed so that's what we've got for you good job sapphire on fixing the reference cooler and we'll link it below if you want to buy one no issues at all recommending the pulse .
when AMD released its navi cards it had a strong proposition in terms of bang for buck when you're looking at raw performance but as I'm sure many of you know the reference caller design was too hot and too loud to be recommended so in a lot of our graphics card reviews recently we've been saying to just wait and see what the third-party Navi card spring and finally we have our first one and it is the sapphire rx 5700 XT pulse card we did also receive the 5700 version of this card but a few testing hiccups have meant that we're saving that review for
later and focus in on the XT model today now in terms of other partners the launch strategy seems to be a bit stagnant we're getting different embargo dates from different partners sapphire was the first to arrive and has the earliest embargo so they're the ones were going with but you can expect to see more reviews very very soon the long story short the Sapphire card is a little bit faster than reference and it is a lot cooler and a lot quieter so it's pretty good news a lot of it is going to depend on the pricing however and that's something that we'll come to at the
end of the review looking at the specs to see what sapphire is doing differently the first change to note is the clock speed now modern graphics cars rely on more of a boost frequency curve rather than exact clock speeds to the figures that you get a more just sort of average reference figures but the game clock which is the important one to look out for AMD cards has gone up by about 3.5% so hopefully we're going to see some higher boosting from this card there's no
memory overclock on the card personally I think this is a shame I do want to see more partners pushing the memory a little bit higher but the pattern does tend to be to just leave the memory at stock these days in terms of how the card actually fits in in the wider market we're still waiting to see clock speeds from other partners so it's unknown at this stage whether this is the sort of average overclock we'll expect or whether it's high or low unlike the reference card
sapphire includes a dual BIOS which now if you move it back to the wards the rear i/o panel you're going to get it into a silent bios mode and this does actually bring the clock speeds down to reference and it has an impact on the fan speeds and we'll look at all that in the performance testing later sapphire has given this card a 241 watt power rating which is 16 watts higher than reference but the power connectors are the same we've got an 8 pin and a 6 pin plug and this
gives a technical maximum available power of 300 watts display outputs are also a match for the reference card we've got three display ports and one HDMI physically the first thing you'll notice probably about the card is that it's very tall it comes up to around 135 millimeters so do watch out for that the good thing is that the power plugs are indented in line with the PCB so you shouldn't have to worry about protruding cables adding even more to the height on the
flipside the length is actually a little bit shorter than reference we measured it to around 260 millimeters they are using a custom PCB to help bring the length down but the other part to note is that it does impede on a third PCIe slot so just watch out for that in your builds the overall build quality is pretty good I'm pretty happy with it you've got a nice strong aluminium backplate all along the back and this is joined by a plastic cooler shroud color wise it's mostly a black and silver affair so mostly color neutral but there are some red highlights and the
Sapphire logo along the top is also red illuminated it's kind of strange these days of RGB to go with one color but I guess it is keeping with the AMD vibe and as far as we know they're gonna have RGB on the higher end nitro cards now the extra height does allow us a fire to install some pretty chunky fans knees on 95 millimeter models and these are going to be expelling most of their air back into the chassis thanks to the open shroud but the horizontal alignment of the fins will mean that summer is going to come straight out of the rear i/o so the first thing I want to
show you quickly is this little quick disconnect fan feature which is something I really like if you just poke a screwdriver between the blades you get access to one screw you can remove and then the fan sort of just pop out of place like so and they have this little connector on the bottom which is pretty handy so if they ever fail it's not like you have to do anything particularly difficult to get a replacement the next step is to take seven screws out of the rear of the PCB and this will allow you to take the backplate and the cooler shroud off so one quick
thing to note is that the back plate does have a little bit of padding on to cool one of the controllers and some of the vrm circuitry and this is something that the reference card did not have so that's kind of good to see with the back plate off you can kind of just lift the main card out of the shroud and there's one cable to disconnect over here note also that the LED draws power from this connector and it has its own connector inside the shroud so if you want to
disable it permanently you can just disconnect it in here flip it back over and you can remove the retention bracket and this will take off the main heat sink whereas AMD was using a vapor chamber for its card sapphire has gone with a more classic heatsink you've got a large copper contact plate and five u-shape heat pipes meandering through a aluminium fin stack note that they're also using thermal paste as opposed to the graphite pad that AMD has on the reference
design last step is to flip the card back over it's four more screws and that will remove this little secondary cooling apparatus this little cooling plate it doesn't connect directly to the main heatsink but it does have some surface area of its own thanks to these little fins it's used to directly call the memory modules via sand pads as well as the MOSFETs of all the power phases around the card so although it's a fire is using a custom PCB the power phase count is the same
you'll see it referred to as seven plus one plus two so the seven plus one part refers to the GPU and it's these seven main phases here and there's this little extra one which is like a secondary phase for the GPU as well and the two is to the memory and that's these two phases here there's also one additional phase right here which is something to do with these of memory
communication interface between the GPU and memory all in all I think Safa is done a pretty decent job with this design the new PCB is pretty compact which is always good to see and most of the main components are directly called as well as there's even a bit of calling on the backplate which is always nice to see as well so we're gonna reassemble this and jump straight into the benchmarks boosting always varies from card to card but larger differences are
unlikely to be overcome by sample variation the AMD reference card tails off fairly quickly as the coulis limitations become apparent but Sapphire is improved calling and overclock give it higher sustained average clock speeds AMD's new boosting is more granular than before hence the jagged lines for all cards but Sapphire has a more consistent pattern regardless of which
bias you use and it's definitely the best card straight out of the box 3d marks granularity makes it good for showing difference between the biases the silent setting is very much in line with reference performance while the default boost bias nudges performance up here it's about 2% but in actual games if 3.5% on average that's not a lot but it's still welcome MD hardware does relatively well in Assassin's Creed Odyssey with performance pretty close to 20 70 Super this is
also where we see the biggest advantage over the stock rx 5700 card at 17% compared to a 14% overall bigger the cards highest framerate is in Far Cry New Dawn and here sapphire is right up there again with 2070 super with a 102 FPS on average the toughest benchmark is Metro Exodus but this is another title where performance is excellent vs. and video a smooth experience here pretty much guarantees solid 1440p gameplay in any title we see great frame
rates again in shadow of the Tomb Raider sapphires card is 13% quicker than our TX 2060 super here although the difference overall is 10% the division 2 is an example of a game that favors Nvidia hardware as evidenced by an overclocked r-tx 2060 super being on par with this overclocked 5700 XT our final title is total war three kingdoms another game that prefers
Nvidia Hardware here gigabytes 2060 super is the faster card but Sapphire still beats this model by 5% over all thanks to the wins that we saw before system power consumption is 16 watts higher using this card versus reference which reflects exactly the difference in total board power and video Hardware is more efficient but it's not as drastic and difference as it once was interestingly the silent BIOS seems more efficient than the reference card despite performance
parity hinting as a slightly more efficient board design overall from sapphire the temperature chart is where sapphires cooler really shows its value a full 10 degrees is knocked off the delta temperature nearly aligning it with Nvidia's founders edition cards comparing the hotspot data shows even bigger advantage for Sapphire but there's no improvement in reported memory or vrm temperatures the silent bios for what it's worth lives up to its name dropping fan speeds about 300 rpm to account for the lower clock speeds this does see a small rise in memory and
vrm temperatures if they receive less airflow - but the core doesn't really get any hotter the regular BIOS is still considerably quieter than the reference card that said you can still stick to the regular BIOS that comes out of the box as this is still considerably quieter than the reference card but you might consider the silent BIOS if low noise is a really top priority overclocking is more involved with AMD than it is with NVIDIA sometimes you'll have a set
overclock that looks good but performance will stay the same or even go down with this card the power limit can be increased by 50% and following that we managed a 50 megahertz overclock in watzmann and got the memory - 15 point 2 gigabits per second this saw performance increased by over 5% in shadow of the Tomb Raider the problem is that it's a hugely inefficient means of getting an overclocked power consumption skyrockets and the
temperature went up by 10 degrees even as fan speeds were approaching 3000 rpm which is way too noisy we'll need to test more cards but unless your water cooling we get the impression that overclocked on the 5700 XT will either be meaningless or too costly in terms of power thermals and noise so there you have it as I said at the start is really going to come down to pricing for this card after all sapphires managed to deliver better boosting definitely got lower temperatures and definitely got better noise but when it comes to pricing we're in a bit of a
different situation depending on the market that you look at starting with the good news we're told to expect a price for this card in the US which is a more than the reference card it's also less than a 2070 Super so it's on pretty good ground considering that it can actually match the Nvidia's card in quite a few titles sadly when it comes to us here in the UK Sapphire seems to be looking into some sort of post brexit apocalypse future and they
attend dollar premium in the US has become a premium here in the UK this car will retail for in it substantially higher than you can get the reference card for now there hasn't really been an explanation offered for this difference and it puts the card cheaper than our TX 2070 super although in fairness that price doesn't really tend to be found anymore the founders Edition for that price is out of stock and
the partner cards tend to go for sapphire does have a bit more room to play with it's just not quite as clear-cut as it is in the u.s. in terms of pricing in the UK you're looking at competition with cards like this gigabyte r-tx 2060 super that we reviewed before now our figures give AMD's card a 5% lead overall but you do have to remember that gigabytes card can be overclocked a lot better and as an Nvidia card it also has the RT cores
which is a feature you just do not get on the AMD Hardware the main thing though is that sapphire really has addressed the cooling of the 5700 XT making it an actually viable choice in the market rather than just a card that helps to bring in review pricing down nevertheless the custom navi story has gotten off to a good start now the thing to remember is that sapphire by
going first makes itself look good because we only have the poor design of the reference cooler to compare to but lucky for you we do have other models already in stock and we'll be bringing you those reviews as soon as we can .
Learn more about Sapphire 11293-01-20G Radeon Pulse RX 5700 Xt 8GB GDDR6 HDMI Graphics Card
GPU Interface PCIe x16
GPU Chipset Radeon RX 5700 XT
GPU Memory Size 8GB
GPU Memory Type GDDR6
GPU Memory Interface 256-Bit
GPU Clock Rate 1670MHz
GPU Boost Clock Rate 1925MHz
DirectX DirectX 12
OpenGL OpenGL 4.5
TDP 241W
Length 254mm
Cooling Type 2 Fans
Power Connectors 1 PCIe 8-pin + 1 PCIe 6-pin
SLI Support
CrossFire Support
DVI-D (Dual Link) 0
DVI-I (Dual Link) 0
DisplayPort 3
HDMI 1
G-SYNC Support
FreeSync Support
the rx 5700 series launched were related to the cooler and the drivers the cooler had serious thermal and noise challenges expected of a blower design and it ended up running unnecessarily hot and loud for a GPU that was otherwise competitive with Nvidia the driver issue has only had partial resolution but overall partner model cards have been expected to solve most of navvies launch issues those relating to thermals and noise and today we're looking at sapphires rx which you saw 100 XT poles card to see if that thermal and noise scenario has
been resolved before that this video is brought to you by Nord VPN Nord VPN is what we use internally as our Virtual Private Network solution we use Nord VPN for international travel to bypass firewalls and at home for data privacy especially from Snoopy and Internet Service Providers Nord VPN helps secure and protect your network privacy and is currently offering 75% off of a three-year plan with one month free claim the offer at Nord VPN comm slash
gamers Nexus and use code gamers Nexus at checkout or click the link in the description below sapphires rx with you 700 XT pulse is more than the reference model that be this card so according to the Cree release pricing we got at more you end up with a cooler that's got a dual fan axial approach it has two 100 millimeter fans so they're large enough that they can spin slower to move more air than a smaller fan and therefore remain quieter and then it's
also got a fairly large aluminum heatsink on a copper cold plate as expected there's a base plate and then the fins are oriented horizontally so about half of your exhaust will go out of the case the other half will remain in the case to be scooped up probably by the CPU fans as opposed to vertically oriented which a lot of coolers are where they go this way and exhaust either up and out towards the glass or down into the motherboard so the sapphire pulse looks good
everything about it looks like a real cooler and a real proper video card design today we're looking at that one we have benchmarks that are really heavily focused on thermal and noise so we've got 40 DB a noise normalized testing versus this card the stock version and then we also have frequency testing so we're looking at the a stable frequency under a heavy workload over a period of time about 30 minutes and that allows us to see does the frequency decay as a result of thermals separately this card the pulse has to be bios's so it's got a switch that's a big
upgrade we liked dual V bios because it gives you an out if you end up breaking something and it also means that there's some modding capabilities that are opened up because again now you have a safety net these second b bios is a slower one the default v bios is towards the front of the card and that's going to be a bit more aggressive fan curve as we'll talk about in the testing and also therefore runs a higher frequency because it can run cooler and the other v bios the non-standard one is quieter it sticks to a much lower fan RPM but also will run into more thermal
issues if you're running in the hotter case so that's the basics of it it's also supposed to be priced instead of the MSRP of the reference card but at ten dollars more expensive for what looks to be a much better cooler and as you'll see in the testing is a much better cooler it's really not that big of a price jump let's get straight into the numbers so we can just talk about how much the cooler really mattered for
this card because that was the big hold point we had on the 5700 XC reference design the reference design is just straight not good we would not recommend it and so sapphire and the other vendors coming out today should start to change the story to be focused more on actual GPU performance and less focused on thermals and noise will go through thermals and noise frequency testing first and then get through the games as always and then we'll have a separate
teardown over the card as well if you'd like to see the cooler design that'll be in a review uploaded on this site as well so subscribe for that let's get into the thermal and noise testing our first chart shows thermals at steady-state after a 30 minute power virus workload measured in a constant ambient temperature of 21 degrees Celsius and monitored every second of the test the Sapphire rx 5700 XT poles clearly solved the reference 5700 x DS biggest
problem outside of the launch driver functionality and that's the problem of thermals with both cards locked to a fixed 40 DBA our noise normalized plots show the thermals for the Sapphire poles GPU edge temperature at 67 degrees Celsius which is a massive twenty seven point one degrees Celsius cooler than reference 5700 XD at the same noise level there noise normalized here they sound the same that's a crazy drop if you ran the 5700 XT under auto settings instead at least with the launch drivers you'd have found a GPU edge temperature of 79 degrees Celsius in exchange for an 11 DBA increase in noise levels decibels are logarithmic and human
perception is hard to objectively measure but in our consultation with sound experts roughly every 10 DBA increased results in a 2x perceived increase in noise to the human ear that's not the same as the increase in acoustic power it's the perceived increase in loudness to a human very different things so just to be extremely clear on that not talking acoustic power GPU Junction temperature which is the governing temperature for throttling and frequency plotted at 110 degrees for the reference card at 40 DBA that's t.j.maxx we're throttling anything
beyond this we can't actually measure because the clocks are now dropping so it would be hotter if we were allowed to run the frequency at the same level the Sapphire card hits 83 degrees giving more Headroom for lower noise levels if you wanted them at any given frequency and temperature that's also a higher frequency for reference even full-auto on the blower cooler the original resulted in 95 degrees for launch memory modules are well under control on the Sapphire card at 40 dB a plot in 82 degrees the reference 5700 X T's 98 degree
result is in questionable territory where G DDR 6 really shouldn't be operating finally MOSFET temperatures around 68 on the sapphire pulse at 40 DB a or 78 on the reference card at an equivalent noise level if you blot the noise level to annoying volumes it's about the same but noise normalizing really matters here these vrm MOSFET numbers are fine for both cards ultimately but Sapphire is doing well at 40 DBA b-roll of the Sapphire card makes it pretty
obvious why it's doing so much better the two fans are big and there's two of them and they're axial so those are all good things so far the Sapphire card is massively improved over referenced for thermals which was our biggest complaint the first one with the reference card a few people out there tried to argue that the reference blower wasn't actually loud or hot but this is why that was wrong it was both objectively hot and objectively loud particularly in the comparative against partner models there's no point in defending a bad cooler design and
Sapphire gives us the numbers to definitively show that there is a better way to make a car at about more than the reference blower it's not even that much more expensive the next chart shows the frequency response of the sapphire rx 5700 XT for each V bios it's got two switches but also compares to the reference 5700 XT AMD card starting with the pulses default V bios the right switch position the Sapphire card holds a range of about 18 80 megahertz the 19 10
megahertz bouncing around hard as a result of aim these various power voltage and thermal thresholds and because it's being pushed pretty high already for comparison the reference 5700 XT by AMD held a range of about 1860 megahertz to 1906 megahertz so they're pretty close to each other with the average closer to 1880 maybe 1885 the reference card bounces around less dramatically but also maintains a lower overall average GPU core clock and for reference the
GPU core clock will change depending on the workload so this won't be the same in every single game or application but we're using the same application tested the same way for each of these cards so they aren't directly comparable finally for the left v bios switch on the poles frequency starts around 1830 to 1850 megahertz but Falls with the rise and thermals to a range of 1815 to 1840 megahertz averaging around 1830 this directly correlates with performance of course so you'd run a lower fps in exchange for the lower frequency but also a lower frequency it means
the card isn't working as hard but left to be BIOS is also less sporadic and more stable in frequency but it's lower performing overall for how this translates to fan speed here's the next chart with just the pulse b bios switch positions the reference card isn't shown since it's a different fan so it direct rpm the RPM comparison isn't linear and doesn't make sense fan speed will be chosen along an RPM temperature curve so HP BIOS has a temperature target then it adjusts the fan to meet the target the pulse start with the default bios tries to maintain
the GPU temperature target of about 71 degrees Celsius edge temperature not Junction temperature during load plotted here the left be BIOS allows more silence and forces a temperature limit about 74 degrees Celsius but it also down clocks so it'd be running hotter here much hotter if maintaining the same 1900 megahertz clock as the first default V bios as for how fan speed is affected the default v bios pegs the pulse fan RPM to about 1570 or just under 40 DBA at 20 inches in a room with a noise floor of 26 DB the left v bios the quieter one pegs
RPM to about 1000 after an initial burst period this heavily controls noise level and brings down the card to about 33 to 34 DBA on an open-air bench at 20 inches and your average computer case you'd never hear this but it also run at least 5 to 10 degrees hotter and an average computer case to be fair we wouldn't recommend this v bios which for any one putting the card under routine heavy loads but it'd be fine for a user with less heavy usage or a case with heavy airflow focus or direct to GPU focus like you know 11 dynamic or something like
that the next question is how the two different view bios has impact PRM and memory thermals we do a lot of stress testing for thermals with fir mark but for this we're just using a constant 3dmark workload that generates something more similar to a game not to be annoying but again remember that this is with lower frequency on the Left V BIOS so the numbers aren't perfectly comparable but this gives us an idea with the default V bios the GPU edge
temperature averages to around 71 degrees in 3d mark or nearing 74 degrees with the other V bios G V Junction temperature the hottest single spot on the GPU is the same in each test roughly 87 degrees this is with ambience accounted for and controlled 221 degrees has a reminder the memory modules show a big difference memory on this card benefits a lot from the fan so droppings 1000 rpm increases the already hot GTD are six modules up to about 92 degrees Celsius in a hot case that'd be entering dangerous territory for memory so we'd again
advise against using this in the hot cases for heavy loads the VRI mosfet temperatures also climb about eight degrees but we're still well within a reasonable operating range Sniper Elite 4 will start us off for games overclocking worked out a bit better with a sapphire card than with our reference 5700 XT but that's because our memory was better on the Sapphire card the actual and the overclocking process is still either broken or just will never have that much Headroom to begin with this is luck of the draw in the memory because the memory on our
5709 xt was also better than our xt but either way we'll take the games full stock the Sapphire 5700 XT pulse ends up at 71 FPS average with the default P bios 60 9.5 FPS average was achieved with the quieter V bios but note that this would be worse in a hotter case and both of these are setting our frame rate as within run to run variants of the 5700 XT reference cards 71 FPS average there is no improvement at stock when it comes to raw frame rate but again
there's a bigger picture of noise and thermals with lower thermals that it give a noise level you can get the same experience for a significantly quieter and that does matter further overclocking the memory worked better and sniper than it did in some of our other games where we were able to hit nine hundred fifty megahertz instead of just nine hundred that we did on the original reference card although in a lot of the other games we tested we did have to
drop down to 940 or a 935 to get stability and everything the 5700 XT pulse at 78 FPS average when overclocked with a low scaling similarly and demonstrating that the overclock is stable in this game 78 FPS average gives us a gain of almost 10% over the stocks sapphires 5700 XT poles or the reference card but this is the last title where we'll see it gain that big from overclocking at 4 and 2018 is next giving us a DirectX 11 classic look rather than the previous dx12 an async compute title at 1080p the RX 5700 xt pulse maintained 151 FPS average which
is within runs run variance of the rx to the 700 XT reference card by AMD lows are also within the wider variants so there's functionally no difference in gaming performance when ignoring thermal acoustic and general quality of life improvements for the Sapphire card toggling into the quieter V bios brought us down to 149 FPS average and overclocking brought us up to 156 FPS average on these sapphire 57 or XT with a 9 50 megahertz memory clock on this title at
1080p although we dropped it at 14:40 and forty a and that resulted in a gain of 3.6 percent over the stock 5700 XT pulse at 1440p the 5700 axial managed to one 15.7 FPS average which is well within the variance of the reference one fourteen point nine FPS average lows were also within variants the quieter v bios Bryn's performs down to one thirteen point six establishing that the default position runs about 1.8 percent faster than the quieter position that said even the default v bios is significantly quieter and cooler than the 5700 AXI reference card so you
don't really even need the quiet one to get an improvement over that overclocking allowed the poles to increase to 121 FPS average a four point seven percent gain entirely thanks to that memory bump at this point we're not sure if the core frequency knob is limited and usefulness broken if na'vi is limited or if the drivers need improvement but we want to see more improvement out of overclocking than we're getting it seems like for the most part this is going to come down to how good is the memory you got on your model and unless any manufacturers
guarantee in a memory clock there's not much you can do about that na'vi really needs that faster memory and seems to follow the trend of AMD cards being memory bound more than anything comparatively in RTX 2070 super is encroached upon by the overclocked 5700 xt cards and the 5700 xt card also performs outperforms the 2060 super which is at 104 FPS average in the stock position at 4k the 5700 xt pulse established that it's fully capable of 4k
gaming with the right gains f1 is one of the easier ones to run although lows begin suffering as we pump the resolution the reference card 68.9 FPS average and the poles at 69 point 4 FPS average are with invariance once again but overclocking bounce the 5700 XD poles it's a 74 FPS average we're ahead of the 2016 Subaru an overclock and stock the stock behind the 2070 super without any tuning the 2070 Super Stock maintains a lead of about 6% over the pulse when both are set to stock settings shadow of the Tomb Raider with DirectX 12 is next 1080p has the
rx100 XT at 124 FPS average for the reference model with the pulse also at about 124 to 125 there's no straight benefit once again other than other quality of life improvements that we discussed earlier it's not like just buying a partner model will immediately improve fps it does depend on the card of course some of them might pre overclock higher than others but ultimately you're looking at a scenario like this where it's about the same out of the box except for the things that we thought really mattered like thermals and noise we dropped the
overclocking numbers at this point - because Shadow the Tomb Raider with any overclock on our pulse at all would trigger a hard shutdown just a black screen it's the same behavior as OCP would trigger which is odd because there wasn't any OCP issues so this is an AMD driver issue and even dropping down to 9 20 megahertz memory caused the same hard crash so we end up giving up on overclocking numbers from this point on and just focused on the basics at the end of the day if all we're doing is overclocking the memory a bit it's kind of the same as
our if it's 200 XT overclocked so there's no point in doing that again what the reference model that is it seems AMD still has a ton to fix and it's broken overclocking behavior for the Navi series cards but at least Sapphire picked up the slack on the cooler at 1440p the rst 700 xt pulse runs at 85 FPS average which is about the same as the 84.5 fps of the reference card the left v bios drops to 83.4 FPS average allowing the default v bios a lead of 2% on the poles card
nothing else here is really worth talking about since it's the same numbers as the reference review aside from noise and thermals there's a clear trend here so no point in showing 4k numbers go check out reference 5700 xt review from launch day if you want the rest of the game benchmark numbers for shadow of the Tomb Raider let's move on strange Brigade with DirectX 12 is up next as a reminder Andy's been doing better in DX 12 the some cards that strange Brigade then with Vulcan at 1440p the 5700 XT Paul stocked the bios frames at 120 4.5 FPS average within variants of the 120 2.8 FPS average of the reference card the left v bios
drops down to 122 FPS average just refined of reference we're going to stop at this point we have numbers for Strange Brigade at 4k with Vulcan we have numbers for hitman 2 and D X 12 and dx11 at 3 resolutions each we have Far Cry 5 and 3 resolutions GTA 5 and 3 resolutions but none of it matters at the end of the day we're all FPS is equal between reference and the pulse so you can check our original review for the rest of those numbers this is all about thermals and noise all that matters for this review other than maybe some follow-ups on
overclocking the game results clearly there's no real difference we have all the other data as stated but if you want the rest of the FPS data you should just check our initial 5700 XT reference card review because the FPS numbers are the same so we're looking for FPS comparison - and video or other AMD cards check that review but for this one it's the same performance as reference so what's really improved is everything else which is important the noise levels you can now run the card at 40 DBA and not incinerate it it's almost 30 degrees cooler with 40 DB a noise normalized the test setup and even if he ran the original one without
that constraint of noise normalizing and a lower fan rpm you'd still be running at like 10 DB a higher than this thing can for the same frequency actually this gets a higher frequency technically even with that noise reduction so Sapphire has done a good job with its pulse for the 5700 XC the cooler is good thermals and noise levels are drastically improved memory temperatures are also dropped to the 80s from nearly 100 degrees Celsius at 40 DBA originally which is a massive step up so if you're buying an RX 5700 XT for sure if you if you've already
decided you're buying one of these cards and it's just a matter of what card do you buy we have no issues recommending the sapphire pulse it's done well it objectively is just massively improved over the original the reference model and we have no issue recommending sapphires version of this cooler we haven't tested the other cooler so of course it's possible unlikely that there are better ones on the market but for what we know today at more than MSRP this one is not only majorly better than the reference design it's probably in the good enough
territory where anything that's improved over this cooler might not be meaningfully so so that's kind of where we stand on it and again we might look at more models later if there's interest in a specific one let us know below and maybe we'll buy one of those now the biggest downside here is still it's on AMD this is not a sapphire issue at this point it's a name the issue and it's the drivers the drivers have been improved so they've updated a few things like that issue with the coloration that we complained about in the original review where the out of box experience for
this thing when we first got it was you install the drivers on any one of I think we tested it on three or four different windows installs and systems and every time without fail after driver install the screen just splashed a really bright green or pink or something and required to restart and it was solvable by restarting it but that's a terrible out of box experience andy has fixed that that's fantastic that's really good news that's an issue that never should have been
there to begin with so it's it's kind of inappropriate to give praise for fixing it good job AMD you fixed an issue that shouldn't have been there and that actually is a pretty big improvement in our books now unfortunately there's still other issues drivers are in a state where they're sort of good enough we didn't encounter any issues when running stock from just one test to the next we were able to run the whole bench suite without any random crashes things like that but there are still issues and most of them are quality of life things like trying to actually change
anything in software with any software not just wot man afterburner as well as this is an AMD api issue but a couple of the issues as quickly as we can recap them I did one take of this already almost 15 minutes long but one of them setting a fan speed does not apply under idle so if you want the fans to spend at forty percent at all times it won't actually start spinning until there's GPU load now we view this as an unexpected behavior for the software if the user sets a fan speed override by nature of being and override the fan should spin at that speed at all times
period so for fan noise test and we actually had to run 3d mark in the background just to change the fan speeds you can't change it on desktop it won't do anything and normally we would do it idle from desktop also there's an issue with reporting sometimes the software report frequencies that are impossible to hit like in the tens of thousands there's also a small bug that's not really too important but at zero rpm mode the fans you'll see them not spinning but softer
reports it as a 4% fan speed like 200 rpm which is obviously not correct setting a higher fans feed is an issue as well you can't bring it back down really so this has been around since Vega frontier edition it is incredibly frustrating that Andy hasn't fixed this yet we have privately complain about it and then we publicly complained about it's still there if you set a fan speed of let's say 70% 80% 90% doesn't matter and you try to bring it down you basically have to
restart to change the fan speed or weight an extremely long period of time we walked away once for half an hour and it reset itself but this has been a bug for a long time it exists an afterburner existed want man it's an API issue or something we want to see it fixed and then also the new driver has black screen hard crash and lock issues that resemble OCP but we don't think it's OCP because that doesn't make any sense this is just one overclocking the Sapphire card this is not a sapphire issue this is an AMD driver issue but basically that problem we run into is when applying even a light overclock of let's say 10 megahertz to memory even
though we know that stable it'll survive eight hours burnin with like a 40 megahertz all increase in 940 megahertz even though we know 10 megahertz extra is stable in some games seemingly randomly we're not quite sure it'll just snap it just shuts down and you have to restart so it's a completely graceless really hard crash and that's bad Andy has this issue where with overclocking not sure if it's just a bug in general or it's some instability that's encountered but and these failures with overclocking when they happen are graceless and so the software will
bug out and be unusable and you'll have to restart or it just shuts down on its own you have to turn it back on or it restarts hard boot so this is an issue where Nvidia does a good job with graceful failures where unless you're under ln2 or pushing extreme clocks for the most part when you're testing stability a clock fails you just restart the application you don't need to restart the whole computer this is something we think AMD can improve upon and we would really like to see it but they seem to be behind still on drivers and on software in some ways
where other parts of their software like the video capture stuff the relive suite that's really good it's just that the poor driver suite needs further improvements overall though we think that they're usable you can play the games some people most probably want to ever notice these issues and then those of you who do we just want you to be aware of them even though you know you can work with them you can live with them it's just kind of annoying quality of life
stuff now let's get back to the roots of this this card fires pulse 50 some 100 XD is very good we have absolutely no issue recommending the pulse cooler sapphire has done a good job with this cooler and it performs well for memory performs well temperatures that is performs well for the MOSFET temperatures performs very well for GPU edge and Junction temperatures so sapphire a job well done if you're buying at 5700 XT at $10 more than MSRP absolutely no
issues recommending this we are happy to say this one's passed our tests now as far as the GPU the thing that AMD makes the 57 rx T the GPU competitively with Nvidia is in a very strong position now the 2060 super is routinely beaten by it the 27 t super tends to hold the lead on average over the 5700 XT few games so that's not true but on average is a bit ahead it's just that that lead is getting small enough now where the price difference doesn't necessarily justify
it so AMD is competing on price as it always does it's in a good spot in terms of price versus Nvidia and for those reasons we can say that the 5700 XC with a partner model not a reference one is a card that we are fine with you purchasing and using for gaming workloads we just ask that you be very aware of all these issues that primarily affect things like again the Fraser is quality of life meaning that it's not we're not running into catastrophic issues that prevent you
from using the device we're just running into issues that are kind of annoying but you can work around them or just ignore them like with overclocking we're eventually like you know what it's 4% if we get it working screw it I'm walking away from this it's not worth any more so you may end up in that boat where it's like you know what is it worth fighting software every day
for 4% maybe no not really so it's it's not like you're losing a ton there but it's something we'd like to see fixed so that's what we've got for you good job sapphire on fixing the reference cooler and we'll link it below if you want to buy one no issues at all recommending the pulse .
Test
System - VGA Rev. 2019.2
|
|
Processor:
|
Intel Core i9-9900K @ 5.0 GHz
(Coffee Lake, 16 MB Cache) |
Motherboard:
|
EVGA Z390 DARK
Intel Z390 |
Memory:
|
16 GB DDR4
@ 3867 MHz 18-19-19-39 |
Storage:
|
2x 960 GB SSD
|
Power
Supply:
|
Seasonic Prime Ultra Titanium 850 W
|
Cooler:
|
Cryorig R1 Universal 2x 140 mm fan
|
Software:
|
Windows 10 Professional 64-bit
Version 1809 (October 2018 Update) |
Drivers:
|
All RX 5700 Series: Radeon 19.7.5 Beta
All other AMD: Radeon 19.5.1 Beta RTX 2080 Super: 431.56 Press Driver RTX 2060S & 2070S: 431.16 Press Driver All other NVIDIA: 430.64 WHQL |
Display:
|
Acer CB240HYKbmjdpr 24" 3840x2160
|
Sapphire 11293-01-20G Radeon Pulse RX 5700 Xt 8GB GDDR6 HDMI Graphics Card review
later and focus in on the XT model today now in terms of other partners the launch strategy seems to be a bit stagnant we're getting different embargo dates from different partners sapphire was the first to arrive and has the earliest embargo so they're the ones were going with but you can expect to see more reviews very very soon the long story short the Sapphire card is a little bit faster than reference and it is a lot cooler and a lot quieter so it's pretty good news a lot of it is going to depend on the pricing however and that's something that we'll come to at the
end of the review looking at the specs to see what sapphire is doing differently the first change to note is the clock speed now modern graphics cars rely on more of a boost frequency curve rather than exact clock speeds to the figures that you get a more just sort of average reference figures but the game clock which is the important one to look out for AMD cards has gone up by about 3.5% so hopefully we're going to see some higher boosting from this card there's no
memory overclock on the card personally I think this is a shame I do want to see more partners pushing the memory a little bit higher but the pattern does tend to be to just leave the memory at stock these days in terms of how the card actually fits in in the wider market we're still waiting to see clock speeds from other partners so it's unknown at this stage whether this is the sort of average overclock we'll expect or whether it's high or low unlike the reference card
sapphire includes a dual BIOS which now if you move it back to the wards the rear i/o panel you're going to get it into a silent bios mode and this does actually bring the clock speeds down to reference and it has an impact on the fan speeds and we'll look at all that in the performance testing later sapphire has given this card a 241 watt power rating which is 16 watts higher than reference but the power connectors are the same we've got an 8 pin and a 6 pin plug and this
gives a technical maximum available power of 300 watts display outputs are also a match for the reference card we've got three display ports and one HDMI physically the first thing you'll notice probably about the card is that it's very tall it comes up to around 135 millimeters so do watch out for that the good thing is that the power plugs are indented in line with the PCB so you shouldn't have to worry about protruding cables adding even more to the height on the
flipside the length is actually a little bit shorter than reference we measured it to around 260 millimeters they are using a custom PCB to help bring the length down but the other part to note is that it does impede on a third PCIe slot so just watch out for that in your builds the overall build quality is pretty good I'm pretty happy with it you've got a nice strong aluminium backplate all along the back and this is joined by a plastic cooler shroud color wise it's mostly a black and silver affair so mostly color neutral but there are some red highlights and the
Sapphire logo along the top is also red illuminated it's kind of strange these days of RGB to go with one color but I guess it is keeping with the AMD vibe and as far as we know they're gonna have RGB on the higher end nitro cards now the extra height does allow us a fire to install some pretty chunky fans knees on 95 millimeter models and these are going to be expelling most of their air back into the chassis thanks to the open shroud but the horizontal alignment of the fins will mean that summer is going to come straight out of the rear i/o so the first thing I want to
show you quickly is this little quick disconnect fan feature which is something I really like if you just poke a screwdriver between the blades you get access to one screw you can remove and then the fan sort of just pop out of place like so and they have this little connector on the bottom which is pretty handy so if they ever fail it's not like you have to do anything particularly difficult to get a replacement the next step is to take seven screws out of the rear of the PCB and this will allow you to take the backplate and the cooler shroud off so one quick
thing to note is that the back plate does have a little bit of padding on to cool one of the controllers and some of the vrm circuitry and this is something that the reference card did not have so that's kind of good to see with the back plate off you can kind of just lift the main card out of the shroud and there's one cable to disconnect over here note also that the LED draws power from this connector and it has its own connector inside the shroud so if you want to
disable it permanently you can just disconnect it in here flip it back over and you can remove the retention bracket and this will take off the main heat sink whereas AMD was using a vapor chamber for its card sapphire has gone with a more classic heatsink you've got a large copper contact plate and five u-shape heat pipes meandering through a aluminium fin stack note that they're also using thermal paste as opposed to the graphite pad that AMD has on the reference
design last step is to flip the card back over it's four more screws and that will remove this little secondary cooling apparatus this little cooling plate it doesn't connect directly to the main heatsink but it does have some surface area of its own thanks to these little fins it's used to directly call the memory modules via sand pads as well as the MOSFETs of all the power phases around the card so although it's a fire is using a custom PCB the power phase count is the same
you'll see it referred to as seven plus one plus two so the seven plus one part refers to the GPU and it's these seven main phases here and there's this little extra one which is like a secondary phase for the GPU as well and the two is to the memory and that's these two phases here there's also one additional phase right here which is something to do with these of memory
communication interface between the GPU and memory all in all I think Safa is done a pretty decent job with this design the new PCB is pretty compact which is always good to see and most of the main components are directly called as well as there's even a bit of calling on the backplate which is always nice to see as well so we're gonna reassemble this and jump straight into the benchmarks boosting always varies from card to card but larger differences are
unlikely to be overcome by sample variation the AMD reference card tails off fairly quickly as the coulis limitations become apparent but Sapphire is improved calling and overclock give it higher sustained average clock speeds AMD's new boosting is more granular than before hence the jagged lines for all cards but Sapphire has a more consistent pattern regardless of which
bias you use and it's definitely the best card straight out of the box 3d marks granularity makes it good for showing difference between the biases the silent setting is very much in line with reference performance while the default boost bias nudges performance up here it's about 2% but in actual games if 3.5% on average that's not a lot but it's still welcome MD hardware does relatively well in Assassin's Creed Odyssey with performance pretty close to 20 70 Super this is
also where we see the biggest advantage over the stock rx 5700 card at 17% compared to a 14% overall bigger the cards highest framerate is in Far Cry New Dawn and here sapphire is right up there again with 2070 super with a 102 FPS on average the toughest benchmark is Metro Exodus but this is another title where performance is excellent vs. and video a smooth experience here pretty much guarantees solid 1440p gameplay in any title we see great frame
rates again in shadow of the Tomb Raider sapphires card is 13% quicker than our TX 2060 super here although the difference overall is 10% the division 2 is an example of a game that favors Nvidia hardware as evidenced by an overclocked r-tx 2060 super being on par with this overclocked 5700 XT our final title is total war three kingdoms another game that prefers
Nvidia Hardware here gigabytes 2060 super is the faster card but Sapphire still beats this model by 5% over all thanks to the wins that we saw before system power consumption is 16 watts higher using this card versus reference which reflects exactly the difference in total board power and video Hardware is more efficient but it's not as drastic and difference as it once was interestingly the silent BIOS seems more efficient than the reference card despite performance
parity hinting as a slightly more efficient board design overall from sapphire the temperature chart is where sapphires cooler really shows its value a full 10 degrees is knocked off the delta temperature nearly aligning it with Nvidia's founders edition cards comparing the hotspot data shows even bigger advantage for Sapphire but there's no improvement in reported memory or vrm temperatures the silent bios for what it's worth lives up to its name dropping fan speeds about 300 rpm to account for the lower clock speeds this does see a small rise in memory and
vrm temperatures if they receive less airflow - but the core doesn't really get any hotter the regular BIOS is still considerably quieter than the reference card that said you can still stick to the regular BIOS that comes out of the box as this is still considerably quieter than the reference card but you might consider the silent BIOS if low noise is a really top priority overclocking is more involved with AMD than it is with NVIDIA sometimes you'll have a set
overclock that looks good but performance will stay the same or even go down with this card the power limit can be increased by 50% and following that we managed a 50 megahertz overclock in watzmann and got the memory - 15 point 2 gigabits per second this saw performance increased by over 5% in shadow of the Tomb Raider the problem is that it's a hugely inefficient means of getting an overclocked power consumption skyrockets and the
temperature went up by 10 degrees even as fan speeds were approaching 3000 rpm which is way too noisy we'll need to test more cards but unless your water cooling we get the impression that overclocked on the 5700 XT will either be meaningless or too costly in terms of power thermals and noise so there you have it as I said at the start is really going to come down to pricing for this card after all sapphires managed to deliver better boosting definitely got lower temperatures and definitely got better noise but when it comes to pricing we're in a bit of a
different situation depending on the market that you look at starting with the good news we're told to expect a price for this card in the US which is a more than the reference card it's also less than a 2070 Super so it's on pretty good ground considering that it can actually match the Nvidia's card in quite a few titles sadly when it comes to us here in the UK Sapphire seems to be looking into some sort of post brexit apocalypse future and they
attend dollar premium in the US has become a premium here in the UK this car will retail for in it substantially higher than you can get the reference card for now there hasn't really been an explanation offered for this difference and it puts the card cheaper than our TX 2070 super although in fairness that price doesn't really tend to be found anymore the founders Edition for that price is out of stock and
the partner cards tend to go for sapphire does have a bit more room to play with it's just not quite as clear-cut as it is in the u.s. in terms of pricing in the UK you're looking at competition with cards like this gigabyte r-tx 2060 super that we reviewed before now our figures give AMD's card a 5% lead overall but you do have to remember that gigabytes card can be overclocked a lot better and as an Nvidia card it also has the RT cores
which is a feature you just do not get on the AMD Hardware the main thing though is that sapphire really has addressed the cooling of the 5700 XT making it an actually viable choice in the market rather than just a card that helps to bring in review pricing down nevertheless the custom navi story has gotten off to a good start now the thing to remember is that sapphire by
going first makes itself look good because we only have the poor design of the reference cooler to compare to but lucky for you we do have other models already in stock and we'll be bringing you those reviews as soon as we can .
Learn more about Sapphire 11293-01-20G Radeon Pulse RX 5700 Xt 8GB GDDR6 HDMI Graphics Card
Specifications
Mfr Part Number: 11293-01-20G
GPU: PULSE RX 5700 Graphics; 2nd Gen 7nm GPU; RDNA Architecture
Engine Clock:
Boost Clock: Up to 1925 MHz
Game Clock: Up to 1815 MHz
Base Clock: 1670 MHz
Game Clock is the expected GPU clock when running typical gaming applications, set to typical TGP (Total Graphics Power). Actual individual game clock results may vary.
Stream Processors: 2560
Memory Size/Bus: 8GB/256 bit GDDR6
Memory Clock: 14 Gbps Effective
Displays: Maximum 4 Displays
Resolution:
HDMI: 4096× 2160@60Hz
DisplayPort: 5120× 2880@60Hz
Interface: PCI-Express 4.0 x16
Output:
1x HDMI
3x DisplayPort 1.4
BIOS
Support: Dual UEFI
API:
Vulkan
OpenGL 4.5
OpenCL 2.0
DirectX 12
Shader Model 5.0
Game Index: Extreme 1440P
Sapphire Features:
Power Design
Fan Check
Free Flow
VR-Friendly
TriXX Supported
TriXX Boost
Intelligent Fan Control IV
Precision Fan Control
Robust VRM Cooling
Robust Memory Cooling
Dual-X Cooling Technology
Tough Metal Backplate
PCB Layers
Two-Ball Bearing
Fan Quick Connect
Fuse Protection
Dual BIOS
AMD
Features:
RDNA Architecture
2nd Gen 7nm GPU
GDDR6 Memory
Power Efficiency
PCI Express 4.0 Support
Video Streaming up to 8K
Display Port 1.4 (HBR3) / DSC
Radeon Software
Radeon Boost
Radeon Image Sharpening
Radeon Anti-Lag
AMD FidelityFx
Async Compute
Radeon Rays Audio + True Audio Next
Radeon FreeSync 2 HDR
Radeon VR Ready Premium
Cooling:
Dual-X Cooling
Dual Fan
Two-Ball Bearing
Form Factor:
2.3 slot, ATX
Dimension: 254 x 135 x 46.5 (mm)
Power Consumption: 241W
OS: Windows 10, Windows 7, 64bit operating system required
System Requirement:
Minimum 650 Watt Power Supply (Recommended).
1 x 8-pin Power Connector.
1 x 6-pin Power Connector.
PCI Express based PC is required with one X16 lane graphics slot available on the motherboard.
Minimum 8GB of system memory. 16GB recommended.
*All images and descriptions are for illustrative purposes only. Visual representation of the products may not be perfectly accurate. Product specification, functions and appearance may vary by models and differ from country to country. All specifications are subject to change without notice. Although we endeavor to present the most precise and comprehensive information at the time of publication, a small number of items may contain typography or photography errors. Products may not be available in all markets. We recommend you to check with sales for exact offers.
Sapphire 11293-01-20G Radeon Pulse RX 5700 Xt 8GB GDDR6 HDMI Graphics Card review
To find out more from Amazon link below
Post a Comment