Asus Radeon RX Vega 64 Strix Gaming (Custom cooler review)

[Elixer]

Hmm...

GeForce GTX 1080 and GTX 1070 finally have some competition, but the release of reference Radeon cards only tells us so much, especially as they struggle with containing noise generated by the thirsty GPUs. What's really needed is add-in board partners to show what the Vega is capable of.

Asus provides a firm glimpse of how RX Vega 64 will perform once in full-retail form, and our opportunity of testing a pre-production Strix Gaming variant has been instructive.

The huge Asus card, using a cooler very similar to that employed on the GeForce GTX 1080/Ti and AMD's own RX 580, is more useful than ever before because it rectifies the problems faced by the reference model.

Performance is up to 10 per cent higher because the cooling enables the underlying RX Vega 64 GPU to maintain better frequencies through all tests, while a trio of larger, slower-spinning fans means the card is very quiet under load, which is an impressive feat considering the power running through it. And as usual, going down the Strix Gaming route leads to understated RGB lighting, FanConnect II, and a change in the default video outputs.

Asus could not have done much more with RX Vega 64 - the card is cool-ish, quiet and about as fast as can be. One worry we have is with respect to pricing; the reference card is already £100 dearer than expected and having it undergo the Asus Strix Gaming treatment is likely to further inflate pricing by, perhaps, another £75. The AMD Radeon RX Vega 64, no matter if it is presented well, really isn't a £600-plus GPU, so we hope that Asus/AMD can work on pricing when this model becomes available in a month's time.

http://hexus.net/tech/reviews/graphics/109078-asus-radeon-rx-vega-64-strix-gaming/
Compare that to:
http://www.guru3d.com/articles-pages/asus-radeon-rog-rx-vega-64-strix-8gb-review,1.html

Final words
You know, I keep thinking, what if. What if AMD had Vega ready like a year ago, battling the GTX 1080. I think it would have been received completely different. People most likely do not frown that much upon the higher power consumption, the performance is there. Where the reference air-cooled cards are simply stated, noisy a custom card like the STRIX does change the and that dynamic a bit. I mean this card is not silent, but silent enough which filters away that extensive concern of mine. Performance wise VEGA 64 STRIX is showing nice yet close to reference numbers. Sometimes a bit more, in others a little less depending on throttling and register variables. With the 8 GB of graphics memory you can game at 2560x1440 really well for a fair amount of money (if you can pick one up at normal prices of course). And hey, it doesn't disappoint in Ultra HD either. There is very little wrong with this GPU overall but granted though, the Vega cards, however, are not a serial game killers right now as it's all a bit average sitting in a performance bracket that the competition offers for the same amount of money for a year now, the GeForce GTX 1080. That, however, is an okay performance bracket to be in, let's be honest here as that is a performance level as that offers a terrific PC gaming experience at say 2560x1440. We do hope to see some driver tweaks that will lift up game performance in the Full HD to WHQD domain a bit more, we still see a bit of CPU overhead with AMD drivers, and that shows at 1920x1080 (it's all very relative though). The brutal truth is that being as close to the reference GTX 1080 Ti really is what this release was in everybody's expectations. If you plan to go for an Ultra HD build, then the Radeon Vega 64 could be a nice match alright. Just connect it to a FreeSync monitor and you'll have a proper gaming experience. ASUS did as much as they could with Vega 64, but that big GPU seems to be a beast to tame on air alright. With this huge cooler the card remains cool-ish (is that even a word?) enough but not by an extensive margin, but it certainly cools better than what the reference products offer. And hey, it is a bit more silent as well. That is a nice improvement overall. Pricing remains trivial, purely based on performance this card cannot exceed the 500~550 Euro pricing bracket as behind that number it just doesn't make sense. But if priced right once volume availability kicks in, we could surely recommend the Radeon RX Vega 64 STRIX over the reference cards - if it fits your budget and brand preference, specifically for the Vega 64. In the end though, the Vega 56 models will make more sense pricing wise relative to performance. But for Vega 64, this pretty much is as good as it is going to get.

I would wait for the other AIBs, and see what they bring to the table.

 

[Reinvented]

Very nice! Can't wait to see what MSI, Gigabyte, and XFX have up their sleeves. Have always loved XFX's coolers so far and performance.

 

[tential]

Also, who adds +15% powerlimit for OCing on an AMD card?

There are too many instances to count where the Vega 64 Strix is underperforming the reference card.
Way too weird.....

This review basically makes Vega AIB and reference seem like there is no delta between them or at least a VERY small one.

That small of a noise delta too?
Edit: HExus review is slightly better, but short, might as well just pick up any Vega card.... not much point waiting for AIB for the 64. Due to the OCing on the 56 and being able to reach approximate 64 levels of performance, that will probably be more interesting to watch for.

 

[Elixer]

That small of a noise delta too?
Edit: HExus review is slightly better, but short, might as well just pick up any Vega card.... not much point waiting for AIB for the 64. Due to the OCing on the 56 and being able to reach approximate 64 levels of performance, that will probably be more interesting to watch for.

Yeah, something is really odd about Guru's review.
I was hoping more AIB cards will make the picture clearer, since I find it hard to believe it didn't perform/sound better that what Guru is showing.

 

[railven]

I was reading the review discussion thread when someone noticed this: I can't verify the actual wording, when I read it earlier today I don't recall the part I put in bold. Reviewer might have added that in or I over saw it but I included the discussion post that pointed it out:
http://www.guru3d.com/articles_pages/asus_radeon_rog_rx_vega_64_strix_8gb_review,4.html

First and foremost, here we have the Vega 10 graphics processor from AMD with its 4096 Shader Cores. I left the thermal compound on there as part of the testing so you can see how it is applied as well as I need to re-assemble the card fast for further testing for a following FCAT review), but all tests have been performed PRIOR to disassembly of the card.

The discussion post:
https://forums.guru3d.com/threads/r...-vega-64-strix-8gb.416698/page-2#post-5469831

Folks are thinking the review is suspect, considering how Hilbert had "left the thermal compound in there as I need to re-assemble the card fast for further testing.".

So two things I took away
1) the issue of the dies not using the mold mentioned earlier is arising, where the whole die+HBM aren't probably making contact is creating issue (as the Guru review mentions the HBM isn't making proper contact)
2) the reviewer gaffed some where and with me not remembering the exact wording and relying on the post, I feel he added in the part I put in bold to clarify all testing was done before removing the heatsink (because if it wasn't you got your answer right there).

The Guru review made the Strix card a meh product, but the other reviews didn't. So something went wrong over at Guru.

EDIT:

Ohh upon going back to get the post from the discussion thread, the reviewer commented on it (rather defensively if you ask me) and made it clear that the pictures were after testing.

His response makes me feel more along the lines the part I put in bold was added after the original article ran to avoid this confusion. The post I was referring to caught it and reading the review as it is written now, it would have never been something to question (which I questioned too because the review results were so bad in my opinion.)

EDIT #2:
Because I'm sick and exhausted, had I read the reviewer's own post in the discussion he admits to re-writing that port of the article. Feel like this should be disclosed as an EDIT because now people reading the revised article wouldn't have noticed. If others, like me and the poster in the discussion thread, assumed the pictures were taken before the testing would have explained the results. Now it just makes the review in it's entirety bad with credibility lost since it was edited without disclosure.

Either way, the review results to jive with other results and reviewer jumping to question a posters intentions for pointing it out are rather dirty.

Bad review all around.

 

[Rannar]

Review was taken offline. It seems card still had default ref. Vega BIOS on it. That's why so small performance delta with reference model.

 

[AtenRa]

ASUS Radeon RX 64 Review taken offline

Yesterday Guru3D Published the Radeon RX 64 STRIX review from ASUS. As everybody had noticed card performs awfully similar towards the reference Radeon RX 64 card. This morning I received a phone call from ASUS, asking us if we’d be willing to take down the article for a few days as they have made a mistake. The sample we received did not get a final BIOS for its final clock frequencies and fan tweaking. Ergo, the sample we received carries a default reference BIOS.

It’s a colossal mistake, but as such the end-results in the review are not representative enough for the final product. ASUS will get the finalized BIOS over once they have finished (likely a day or two) after which we will re-test the card with that final BIOS and thus republish the review. All this explains why the STRIX card was so incredibly close to Vega 64 performance. Apologies for the inconvenience, but this mistake was not one coming from us.

 

[Shmee]

Hmm, I see. Looking forward to more custom cards, I expect the Nitro series could be great.

 

[Krteq]

ASUS Radeon RX 64 Review taken offline

Huh, how can ASUS f* that so much? Waiting for retest

 

[railven]

Really is embarrassing for Guru if you ask me. Surely they've had noticed the clocks weren't hitting what ASUS probably stated in the adverts. More so, I recall the article opening up with something such as "this is retail version unlike other reviews that had preview samples" or some nonsense to add authority to their review.

Also remembering the page from the article with clocks, the base and boost clocks were basically the same for the stock and strix in the table they posted. That should have been a red flag for them.

Time to see what a proper BIOS gives for results.

 

[eRacer]

Really is embarrassing for Guru if you ask me. Surely they've had noticed the clocks weren't hitting what ASUS probably stated in the adverts. More so, I recall the article opening up with something such as "this is retail version unlike other reviews that had preview samples" or some nonsense to add authority to their review.

Also remembering the page from the article with clocks, the base and boost clocks were basically the same for the stock and strix in the table they posted. That should have been a red flag for them.

Time to see what a proper BIOS gives for results.

Guru3D didn't do anything wrong, other than perhaps giving this one particular AMD-based card a bit more praise than it deserved. ASUS is the one the should be most embarrassed for the BIOS mistake, and to a lesser extent AMD for various limitations that have hobbled the Vega roll out.

Other Strix previews showed the card using the same base and boost clocks as the reference card, so why should Guru3D assume anything different? Due to the power hungry Vega 64 architecture the Strix card has little chance of maintaining the reference max boost of 1630 MHz at a tolerable level of power consumption. The Computerbase Strix preview showed the Strix card settling around 1500 MHz under full load. I'm guessing final Strix performance will be similar to that of the other previews.

 

[railven]

Guru3D didn't do anything wrong, other than perhaps giving this one particular AMD-based card a bit more praise than it deserved.

That's kind of bad in itself. It was basically no different then a reference Vega 64, and they just assumed that was fine. The data itself should have been the red flag.

ASUS is the one the should be most embarrassed for the BIOS mistake, and to a lesser extent AMD for various limitations that have hobbled the Vega roll out.

ASUS should be embarassed too. But swing by the reddit thread over at r/AMD. It seems lots of people caught the original article before the edit. The original article's wording made it easy for someone to infer that the cooler was taken apart before the testing was performed. Lots of people started to use that as the excuse for the poor data results. Guru didn't even disclose it in the article that they had reworded it and you'd have only known if you read the thread.

EDIT: Here is the reddit thread:
https://www.reddit.com/r/Amd/comments/6yls5j/guru3ds_asus_radeon_rx_64_review_taken_offline/

When I read the original article the data to me was off. I also assumed due to the poor wording he had done testing AFTER disassembly. So I wrote the article as bad from my first go through. I peaked the discussion thread to see if anyone else thought similar. One person did. And he got accused of being an AMD shill for calling to question the poor article's wording and data.

Other Strix previews showed the card using the same base and boost clocks as the reference card, so why should Guru3D assume anything different? Due to the power hungry Vega 64 architecture the Strix card has little chance of maintaining the reference max boost of 1630 MHz at a tolerable level of power consumption. The Computerbase Strix preview showed the Strix card settling around 1500 MHz under full load. I'm guessing final Strix performance will be similar to that of the other previews.

That's just it. You might feel I'm coming off hard on Guru, but they opened the article stating this was a retail card, ie adding authority over those other articles that had preview samples. Only for it to basically end up as a giant mistake. Hits their credibility in my opinion.


Just ask yourself, why does a custom AIB have barely any improvements over a stock Vega 64 card even if the cooler in itself should at least reduce temps and noise by a considerable margin. The article was bad from the moment you got into the data and then the conclusion just made it even worse. I'd have called that card a spade.

 

[Elixer]

That's kind of bad in itself. It was basically no different then a reference Vega 64, and they just assumed that was fine. The data itself should have been the red flag.

Yeah, if the numbers don't make sense, then a good reviewer would have gotten in touch with ASUS and ask them if they care to comment on the results they have so far.

That is mainly the reason I put Hexus's preview up as well, it is like two completely different cards.

 

[Rifter]

Guru3d lost alot of my confidence after this. Ill be looking else wear for reviews.

How can you trust them to give you relevant and correct information and make informed reviews of products when the reviewers are to stupid to notice that the AIB card performs the same as the ref card? something that never happens.

Like wow. Most of us would have noticed this let alone people who do this for a living, i just dont understand how you look at the data and go yup looks good to me.

Their credibility deserves to take a huge hit.

 

[railven]

I wouldn't go so far as to rule out the site, all data is welcomed, but bad data should be tossed out and challenged. I guess I'm more salty because of how the reviewer handled the criticism someone brought to him. Comes of as uber douche-like especially when it turns out they got a bad card. Even in his response he throws a jab at the preview articles as if they Guru's article is the definitive article.

https://forums.guru3d.com/threads/r...-vega-64-strix-8gb.416698/page-2#post-5469841

Whenever people see something they do not like it is always the media outlet that is suspect. BTW your IP traced back awfully close to the AMD Markham Canada HQ? Is that a well educated observation to make or an incredible coincidence?

1) The disassembly photo-shoot of the card was done AFTER all performance tests had been done, ergo the card is tested in it's default state. The remark was made due to the fact i still needed to use the card for pending FCAT tests. I also make the photos like that so that people can see how TIM is applied. But I'll rewrite that a bit to make it more clear.

2) We tested in a final configuration. All other previews you refer to are just that, based on preview / non-final early sample cards, some media had quick access to a card for merely an hour or so as ASUS dropped by to show them. No-body tested with long duration temps and DBa measurements with the card properly warmed up, we did. Other future media reviews will back the temps we are seeing, unless our sample had an isolated problem of course.

3) The performance differential in-between reference and the ASUS card is 100% based on throttling, and VEGA64 sure is throttling up and down a lot causing small FPS differences.

But you know what - I hope it is an isolated issue on our side, I honestly doubt it, but I do hope it. But be glad we do not cover up our findings and post results as they are measured here in the lab.

Person challenged the data, author responds by accusing him of being an AMD shill. That's just bad form.

 

[Rifter]

I wouldn't go so far as to rule out the site, all data is welcomed, but bad data should be tossed out and challenged. I guess I'm more salty because of how the reviewer handled the criticism someone brought to him. Comes of as uber douche-like especially when it turns out they got a bad card. Even in his response he throws a jab at the preview articles as if they Guru's article is the definitive article.

https://forums.guru3d.com/threads/r...-vega-64-strix-8gb.416698/page-2#post-5469841


Person challenged the data, author responds by accusing him of being an AMD shill. That's just bad form.

Wow, that response is even worse, these guys just cant stop from digging themselves deeper. This just went from lack of intelligence to notice the data to jumping the shark entirely.

 

[kawi6rr]

I'll pass judgment on AIB cards when I see Saphire, XFX, and MSI review's for a few reputable sites.

 

[amenx]

Hilbert Hagedoorn, the owner of Guru3d is a prolific reviewer. He has literally reviewed hundreds of GPUs over the years (since 1999/2000?) as well as CPUs, motherboards, SSDs, etc so can imagine him to be quite a busy fellow. I generally respect his reviews, just that this last one was out of character. I might put it down to being too busy to have focused too much on the difference other reviewers had with their samples or on carefully crafted responses to these differences. As an active member in his forum, and very familiar with his work, I tend to trust his integrity. He also does do updates to reviews when something needs to be corrected or re-examined, but have always seen him mark these as 'updates'. So again, out of character if he has not done that here.

 

[Mopetar]

I think part of the reason the reviewer was so agitated is because prior to the actual issue being identified a lot of people were jumping to the conclusion that he benchmarked it after pulling the cooler of to get shots of the thermal paste. Even after he updated to clarify this, there were still some tinfoil hat types that were thinking he was trying to do a coverup.

As to why he wasn't as suspicious of the results as he maybe should have been I don't know. I can see it being an easy mistake though as you wouldn't expect ASUS to have made the mistake they did. Also if we're being truly honest Vega launch has been enough of a mess already that it could have been easily written off as part of that fiasco.

Everyone should just calm their tits and stop getting worked up over it. The review was pulled after the BIOS issue was discovered and it's likely that we'll get the actual results soon enough.

 

[railven]

I think part of the reason the reviewer was so agitated is because prior to the actual issue being identified a lot of people were jumping to the conclusion that he benchmarked it after pulling the cooler of to get shots of the thermal paste. Even after he updated to clarify this, there were still some tinfoil hat types that were thinking he was trying to do a coverup.

As to why he wasn't as suspicious of the results as he maybe should have been I don't know. I can see it being an easy mistake though as you wouldn't expect ASUS to have made the mistake they did. Also if we're being truly honest Vega launch has been enough of a mess already that it could have been easily written off as part of that fiasco.

Everyone should just calm their tits and stop getting worked up over it. The review was pulled after the BIOS issue was discovered and it's likely that we'll get the actual results soon enough.

I don't think anyone here is actually worked up. I just posted my own findings from the article. What I put in bold in your post is no one else's fault but the author's. He worded it so poorly in the original run anyone reading it could easily assume he ran the tests after taking it apart and stating he didn't reapply any TIM. (Why even state this to begin with? You're asking for confusion.)

I think the reason people are even throwing this review under the bus is because the Vega launch was such a blunder. It's like even here, where AMD can shine, the first official review (and trust me the author stated it) on an custom card is basically garbage.

 

[96Firebird]

I would be pissed too if I wasted my time benchmarking a card to get meaningless data because it doesn't represent a retail card.

 

[Mopetar]

I don't think anyone here is actually worked up. I just posted my own findings from the article.

It wasn't bad here, but there were other forums where people were acting like poo flinging chimps.

I would be pissed too if I wasted my time benchmarking a card to get meaningless data because it doesn't represent a retail card.

Yeah, reviewers are typically always busy with something and they were one of the few sites that had a review which typically means more traffic. It's just a crummy situation to be in all around.

 

[richaron]

Posters in this thread are obviously not scientists.

When running experiments (aka benchmarking in this situation), you don't let your expectations cloud your results. You report the numbers you get. Sure he might have thought something was up, and maybe was in a hurry, but he reported the numbers he got. People don't like reading over boring experiments, or those which don't provide cohesive or decisive data. But it's all part of the grind.

 

[JoeRambo]

Also remembering the page from the article with clocks, the base and boost clocks were basically the same for the stock and strix in the table they posted. That should have been a red flag for them.

The irony is that at the moment it is impossible to find those "base and boost" clocks. Not even on Asus site. But people are stupid for sure if they are hoping that AIB air card will clock better/higher than in between Vega air/WC editions. And expecting reviewer to notice and dig when actual full difference between those two is maybe 10%? Wow.