I was talking about Youtube also. PC Perspectives and Gamers Nexus both ran review summaries on YouTube.
They both also have website based write up of their reviews.
The PC Perspectives LiveStream Review on Youtube is here:
Radeon Vega Frontier Edition LIVE Benchmarking - YouTube
The Gamers Nexus Reviews are here:
I also found this review which compares the Vega FE Edition to a GTX 1080Ti
AMD Vega FE Overclocked Gaming Benchmarks - YouTube
OK, so in any case most of the Youtube "reviews" or "comments" from "single guys on Youtube" refer to the PCPer/Gamers Nexus Reviews or #define guy who got the card first. And it does not seem that there is much disagreement between them all regarding the overall performance of the Vega FE when running games.
The problem is they all show that the Vega FE is performing somewhere between a reference GTX 1070 and 1080 on games.
The drivers used are the latest available from AMD for the Vega FE.
The analysis of the voltage regulator on the Vega FE looks like it is very high quality, so an aftermarket card is unlikely to help much there. Maybe the HBM2 Vreg might be able to be improved ...
Although the card is a blower type, again it does seem to be very well constructed, as it should be for $1000.
Perhaps an internal 3 fan gaming cooler with a large vapour chamber and radiator cooling will help lots, but will it help improve gaming / firestrike performance by 30-40%?
Maybe the Vega Chip needs watercooling?
Since the Vega FE does seem to beat the Nvidia Titan-Xp in some places in the "Radeon-Pro" software space, it looks to me like the raw compute power is there.
But remember these drivers are not Certified.
Perhaps some new architecture feature of the RX Vega is not yet working when running DX11/DX12/Vulkan based Games.
Setting the Vega FE driver in "gaming mode" just changes the Crimson Interface, not anything in the driver behaviour, according to reviewers who have been in touch with AMD to get information about it.
To say that the Vega FE is "Not a gaming card" when it has a "gaming mode" in the driver is not really a good enough explanation for some.
It does look to me like the Vega FE and Vega Chip is simply underperforming and hopefully that is because the the RX Drivers for gaming are not ready, at least on the Vega FE card.
The fact that it is only 1 month before the RX Vega is supposed to be launched/announced does not seem great.
I can remember that the AMD Crimson ReLive Driver, although it was a great improvement, took around 3 months after launch to work reliably enough to use on my System.
If the gaming drivers for RX Vega are not ready then I think that releasing the Vega FE like this, even then though there was emphasis by AMD that this is not a card for gamers although you can run games on it, is a problem. Foe example I watched a review with AMD staff present in the review where they did a quick demo of Doom running on a Vega FE but they did not want the reviewer to show the game FPS!
I think that AMD should try to properly address why the Vega FE Gaming performance is only between GTX 1070-1080 performance as soon as possible.
Another month or two of more and more "dissapointing" gaming reviews done on the Vega FE before the RX Vega is released and available to reviewers is not good.
I am really hoping that the RX Vega Cards perform better than the Vega FE in gaming and I am also hoping that the gaming performance of the
Vega FE gets seriously improved and soon.
I am in the market to purchase either a Titan Xp or a Vega FE/ High end RX Vega. And I have been "waiting for Vega" for almost too long now.
So I have the FE's (had to get them for a test project) and my numbers are pretty much in line with what the other reviewers came up with. I expected these numbers for the most part, especially when it comes to the gaming side of things. Then again, I also did not get these cards to game with them.
One site I do not like to reference is "Videocardz", but so far their pre-release benchmarks have mostly been accurate and they've put a table up on what seems to be RX Vega, putting them in line with factory overclocked 1080's. Am I hyped up? You betcha - all I need now is my cooling manufacturer (EKWB) to come up with waterblocks with the next 2 months and I'll be a happy camper!
image and benchmark source: VideoCardz
Thanks for those results,
What is the difference between the #1 to #9 runs then?
They all show 1620MHz/1890Mhz.
Are they the Radeon Crimson Global/Profile Settings from default?
Are they all different cards?
I guess I will go and look at that site.
In any case they all show numbers between
a 1070 ~ $500
a 1080 ~ $650.
all on air, not watercooled.
That's one thing I could never figure out with their tests - they don't indicate the "+" values next to the core clock, so I'm assuming it's a core OC. All I care about is that it's well above my current stable OC settings of 1510 CC on my 480's (Oscar Mike )!
I don't know what to make of this article ... but here goes ...
I really, really, hope it is true.
But still I think something needs to be done from AMD to address the Vega FE gaming performance seen by Gamers Nexus/ PC Perspectives.
Maybe work with them and let them test early RX Vega driver based improvements in the Vega FE Drivers?
I found this review interesting. How does the Vega FE compare to the R9 Fury X when run at the same GPU Clock speeds in "Radeon-Pro" applications and in Gaming.
It seems to suggest to me that the Vega FE is essentially running Fury X drivers for the games at present, and possibly not exploiting any of the new Vega Architecture features in the gaming tests so far.
The Vega FE gaming performance is pretty much a tie with the Fury X when both GPU's are are clocked at the same speed during gaming.
The Vega FE outperforms the Fury X cards significantly in "Radeon-Pro" applications when both cards are clocked at the same speeds.