Friday, June 30th 2017
New Performance Benchmarks of AMD's Vega Frontier Edition Surface
You probably took a long, hard read at our article covering a single-minded user's experience of his new Vega Frontier Edition. Now, courtesy of PCPer, and charitable soul Ekin at Linus Tech Tips, we have some more performance benchmarks of AMD's latest (non gaming specific) graphics card.
Starting with 2560x1440, let's begin with the good news: in what seems to be the best performance scenario we've seen until now, the Vega Frontier Edition stands extremely close to NVIDIA's GTX 1080 Ti video card in Fallout 4. It trails it for about 10 FPS most of the test, and even surpasses it at some points. These numbers should be taken with a grain of salt regarding the RX Vega consumer cards: performance on those models will probably be higher than the Frontier Edition's results. And for the sake of AMD, they better be, because in all other tests, the Frontier Edition somewhat disappoints. It's beaten by NVIDIA's GTX 1070 in Grand Theft Auto V, mirrors its performance in The Witcher 3, and delivers slightly higher performance than the GTX 1070 on Hitman and Dirt Rally (albeit lower than the GTX 1080.)At 4K (3840x2160), the Vega FE trails the GTX 1080 by about 3 FPS (at 57 FPS, just shy of 60 FPS) on Dirt Rally; trails it again (this time with a 7 FPS difference) in Fallout 4, at around 42 FPS; delivers around 66% of the GTX 1080's performance on Grand Theft Auto V, and less than 50% of the GTX 1080 Ti's performance on the same game. In Hitman, the Vega FE delivers around 83% of the GTX 1080's performance (around 50 FPS versus the 1080's 60), and delivers almost the same result on The Witcher 3, barely maintaining a 30 FPS performance towards the end of the run.Do note that all of these tests will apparently be re-run by PCPer, and the publication is looking to publish their results later today. Also keep in mind the Vega Frontier Edition isn't a consumer graphics card, and isn't officially meant for gaming. Instead, it's meant for professionals or prosumers who do some professional workloads as well as some gaming, and want to have the ability to test their development fruits with the same graphics card they developed with. Power draw was rated at around 280 W while gaming, with only 25 of those being taken from the PCIe slot, which seems somewhat disproportionate.
Apparently, there was some testing done on mining software as well, and performance is reported as disappointing (as in, "very low".) This probably speaks to the HBC (High Bandwidth Cache) and HBCC (high Bandwidth Cache Controller), which probably will require a lot of fine tuning from mining software (remember the GTX 1080 is generally poor in mining workloads compared to the GTX 1070 because of the higher latency of its GDDR5X memory implementation.) Perhaps these news come as a relief, however, since availability of RX Vega cards will likely be limited without miners taking up the whole supply.
Sources:
Ekin @ Linus Tech Tips, PC Perspective, PC Perspective YouTube Channel
Starting with 2560x1440, let's begin with the good news: in what seems to be the best performance scenario we've seen until now, the Vega Frontier Edition stands extremely close to NVIDIA's GTX 1080 Ti video card in Fallout 4. It trails it for about 10 FPS most of the test, and even surpasses it at some points. These numbers should be taken with a grain of salt regarding the RX Vega consumer cards: performance on those models will probably be higher than the Frontier Edition's results. And for the sake of AMD, they better be, because in all other tests, the Frontier Edition somewhat disappoints. It's beaten by NVIDIA's GTX 1070 in Grand Theft Auto V, mirrors its performance in The Witcher 3, and delivers slightly higher performance than the GTX 1070 on Hitman and Dirt Rally (albeit lower than the GTX 1080.)At 4K (3840x2160), the Vega FE trails the GTX 1080 by about 3 FPS (at 57 FPS, just shy of 60 FPS) on Dirt Rally; trails it again (this time with a 7 FPS difference) in Fallout 4, at around 42 FPS; delivers around 66% of the GTX 1080's performance on Grand Theft Auto V, and less than 50% of the GTX 1080 Ti's performance on the same game. In Hitman, the Vega FE delivers around 83% of the GTX 1080's performance (around 50 FPS versus the 1080's 60), and delivers almost the same result on The Witcher 3, barely maintaining a 30 FPS performance towards the end of the run.Do note that all of these tests will apparently be re-run by PCPer, and the publication is looking to publish their results later today. Also keep in mind the Vega Frontier Edition isn't a consumer graphics card, and isn't officially meant for gaming. Instead, it's meant for professionals or prosumers who do some professional workloads as well as some gaming, and want to have the ability to test their development fruits with the same graphics card they developed with. Power draw was rated at around 280 W while gaming, with only 25 of those being taken from the PCIe slot, which seems somewhat disproportionate.
Apparently, there was some testing done on mining software as well, and performance is reported as disappointing (as in, "very low".) This probably speaks to the HBC (High Bandwidth Cache) and HBCC (high Bandwidth Cache Controller), which probably will require a lot of fine tuning from mining software (remember the GTX 1080 is generally poor in mining workloads compared to the GTX 1070 because of the higher latency of its GDDR5X memory implementation.) Perhaps these news come as a relief, however, since availability of RX Vega cards will likely be limited without miners taking up the whole supply.
166 Comments on New Performance Benchmarks of AMD's Vega Frontier Edition Surface
www.gamersnexus.net/hwreviews/2973-amd-vega-frontier-edition-reviewed-too-soon-to-call
Seems to confirming pcper's position of card on current driver stack: gaming performance between gtx0170 and gtx1080.
AMD delayed Vega RX as much as it could. I mean, there where half a dozen events where they could introduce this card, and they chose Siggraph that has nothing to do with gaming? Obviously the choice had to do with dates. Siggraph was the last one on the list. So, it seems they are not ready yet. In my opinion, if someone was benchmarking an RX Vega TODAY, the performance of RX Vega would have been about the same as FE, if they where running at the same frequency. If RX Vega hits 1700MHz stable, then I guess we will be looking at a power hungry card that would offer at least performance somewhere between 1080 and 1080Ti. If there are some features in the cards, both RX and FE, that are not yet enabled, features that could offer higher performance, then and only then we could see a 1080 Ti competitor, and why not a 1080 Ti killer.
The market needs a 1080 Ti killer. Nvidia fans need a 1080 Ti killer, more than anyone else. Nvidia fans not having a modern Nvidia card need to go to the latest models, because we all know - or at least heard that ugly rumor - about how old Nvidia cards age badly. Nvidia fans that do have a good modern Nvidia card, also need a 1080 Ti killer, because not only Nvidia cards, but also GSync monitor prices are staying high, if not increasing. I was looking at the latest 34UC89G-B from LG. $1000. The equivalent 34UC79G with FreeSync, that was out in last September, but have almost identical specs, had a starting price of $700. We are talking about $300 difference. And considering that now sells for under $600, the difference is over $400. It's only one model but it seems that as long as AMD doesn't have hi end gaming cards, Nvidia fans needing a hi end GSync monitor, would have to pay much more for that GSync tech, compared to the past. Everyone NEEDS a 1080 Ti killer.
If you call it a professional card - fine. But it's not very good at professional tasks as well - often loosing significantly to a "pure gaming" (your words!) Titan Xp.
And it uses more power, runs at higher temperature and make more noise.
Is there anything that this card does well? :-)
(other than being better than previous AMD offerings, so it should win hearts of die-hard AMD fans...)
Low-level scheduling is after all not controlled by the driver. The GPU itself controls assigning the workloads to various GPU clusters, fetch data/textures, etc. The driver only sees queues of operations, while the GPU scheduler balanced the load, handles resource dependencies/hazards, etc. It is in fact kind of analogous to what your OS kernel sees in the CPU; it sends a chain of instructions, which the CPU's prefetcher decodes, executes out of order, optimizes, prefetches, guesses branches, etc. You can never do this sort of stuff in software, since this has to be completed in clock cycles, which means on a ns scale. The same is true for GPUs, it's just even more sensitive there, if they were to be tightly managed by the CPU the overhead would be huge. If something doesn't work right in the hardware, then the driver can't fix this, it can only enable/disable feature sets, and of course not in real time. It's called PR bullshit, and everyone does it. The only thing which will happen between now and RX Vega is minor driver tweaks, and of course AMD trying to get a couple of game developers to "optimize" for their hardware so they can claim it shows the "true potential". Exactly, Vega FE would be useless for development if it performed vastly different from consumer cards. Yes, and considering Volta will arrive next year and Navi will arrive roughly a year after that, AMD needs to at least compete well with the upcoming GV104 to even be relevant.
I've never seen such abundant ignorance when NVIDIA had anything delayed or released in different than usual manner. But for Vega, everyone is freaking out with such illogical BS it's making my head hurt real badly.
Once again: what is this card good at? Why would anyone choose it over a Titan Xp?
From pro.radeon.com/en-us/product/radeon-vega-frontier-edition/
"The Radeon™ Vega Frontier Edition graphics card is designed to simplify and accelerate game creation by providing a single GPU that is optimized for every stage of this workflow, from asset production, to playtesting, to performance optimization."
So I'm not really into game development or even hardcore gaming, but wouldn't it make sense for game-development GPU to perform similarly to the best gaming one from same generation? Especially if it's meant to be used for "playtesting" and "performance optimization".
Also, after going through the Vega description, I have to point something out.
I've been following PC stuff development for quite a while now (possibly longer than many TPU forum users live) and I don't think I've ever seen such a marketing BS. :-/
"
Be first. Be the Pioneer.
The Radeon™ Vega Frontier Edition Story
Who are the pioneers? They are the ones who have cured diseases and strengthened our bodies. They work to heal our planet and explore new ones. They work to undo mankind’s mistakes and protect the next generation from making them again. Harnessing science to fuel creativity, and employing creativity to drive science. They pursue an unerring, unwavering path towards their goals. There are no barriers, no compromises. They are people who see boundaries as starting lines, and who risk everything in pursuit of innovation. They are the early adopters, the people whose passion is to pursue what is new and different. Their achievements won’t be measured in days, weeks or even years. They’ll be measured in centuries.
"
2. Assuming that a contencreator using etc. maya performance is will get preportional performance ingame is like saying that render perf should be equal to gaming performance. (Im pretty sure that you agree that that is not the case)
3. It may be possible that the RX Vega is 1:1 with the FE (but we dont know), so presenting that as a fact (as alot of people do), based off of their possitive/negativeness towards AMD is not a solid argument for the performance. There is so many unknowns that people are ignoring (again guessing based off of their own "knowledge".
4. We can all agree that a 1080Ti killer would be nice but I would argue that that is not what VEGA is. VEGA is a solid step into a WAY longer strategy using HBM2 new chip design new (almost everything) What we need is a card that can perform like a 1080+ but at a competive price. Right now I run all games on my 1440p monitor with a 1070 (and it will be enough for 95% of the gamers out there). The problem is that not 95% wanna spent that much money on a 1070+ card let alone a 1440 144 hz monitor so they are the 95% under it. THATS what AMD is targeting. with 460/470/480 they made a huge push to Nvidia's pricing, with VEGA they should hopefully force Nvidia to do the same again, but Nvidia can push out binned 1080Ti-chips out and with a few improvements win the crown, but the crown is the 1% the rest is where the money is. A RX VEGA will push a 1440p in current games to the needs IMO. The 4K monitors with either gsync/freesync is still a 95%+ thing (ALOT of people are still sitting on a 1080 60 hz screen.) Would I like RX VEGA to beat 1080Ti? sure but its not what its for. Its for reclaiming market and pushing new technology; kinda like Ryzen is; it doesnt win the "gamingcrown" but still it is selling like butter to almost every regment due to price performance.
Your trolling and assuming that Vega FE will be representative of RX Vega in gaming is either plain ignorance or fanboyism/AMD hatred or a mixture of both. Do you not think that AMD know's exactly how FE and RX will perform compared to each other yet still decided to release the FE knowing it's gaming performance is lacking, hang on scratch that, why would they develop Vega at all for 2+ years and $$$$$'s of R+D costs knowing that based on FE gaming results it's no better than Fury X, heck even if they tweaked polaris with higher clocks/TDP it would be on the same level, so why didn't they just do that? Not too mention you don't even know RX Vega release specs/clocks etc but hey why let facts or reasoning get in the way of a good ol' fashioned fanboy witchhunt :kookoo::nutkick:
Raja saying that the FE card is not for gaming, doesn't mean that FE is bad in gaming. Just that there is no reason to go and pay for extra pro features in drivers, that are useless in games. He is just more honest.
Also, no one says that those two cards will be 100% identical. RX will have probably higher frequencies and maybe better cooling solution. I wouldn't be surprised if the RX is at the same level of performance, or a little better if clocked higher, compared to the liquid version of FE. Now, if AMD manages also to enable some features that are now disabled in the next days, features that increase performance in games, those features will become available to the FE card also, improving it's performance in games. You'll see that, if it happens.
2. I am not sure what you try to say here, but no one is judging a card's gaming performance based on the performance it has in pro applications. Also no one comes to conclusions about gaming performance, based on the performance in pro applications. So your example here is wrong. No one cambares oranges with apples.
3. If you don't like an assumption, that doesn't mean that people who believe that this assumption is closer to the truth, are negative to AMD. I would not point you to my system specs(3 AMD PCs) to see if I am positive or negative towards AMD, I would just suggest to you to look my older posts. Believe me. You are not a bigger AMD fan than me.
4. I was always saying that Vega was meant to fight 1080 and beat it. But Vega was delayed and every time Nvidia was coming out wth a new high end GPU, people where assuming that Vega will go against that new GPU. Not AMD's fault. On the other hand AMD enjoyed this speculation and never came out to clarify which card from the competition was it's target.
As for the rest you write, I will only say one thing. Having the fastest cards, helps to sell many mid/low end cards, even when the competition is offering better models. No it's not a gaming card, it's just a card that offers a gaming mode and also is targeting developers of games. How could I misunderstood that?
Thank you for your wisdom.