Apparently the fucktards at AMD decide they should go full fucktard with this card. No partner models and only 5K being made total, those are the rumors but everything is pointing to more disappointment. Navi? Who knows… who knows if it will ever come out. I’m guessing if it actually is going to be made (which with the chaos of AMD’s GPU division they could scrap it for the sheer hell of it) there will be another keynote next year at CES with the announcement, maybe. That V64 will serve you long time, well into Navi.
Going by spec, Vega Vii is MI50 that failed the tests, waste product so to speak.
AMD competing in the high-end GPU segment at all is a bad move historically and financially.
When Nvidia does that, noone cares.
That is BS. Return on investment would be so negative, share holders would grab pitchforks and torches over that decision.
There is competition to the 2080 for 2080 price. I fail to see the disappointment.
4 Likes
If its true only about 5K (which I could see given no partner models) its a disappointment.
2080 having no partner models would get them called out. This is not about NVidia though. This V7 fixin to be another mess of the latest AMD messes. I hope I’m wrong, I hope they make enough to satisfy demand and sell but being they are just re-purposed parts the supply won’t be that great.
I getting the feeling its a collectors item. If it’s at MSRP and techtube reviews benchmark it as 2080 level performance. It would be a good card for many uses. Considering Vram and memory speed.
It is an affordable compute card that happens to do gaming. Some of the numbers I have seen are promosing, put the price gap compared to Vega leads me to believe they have a cut down version in the making.
AMD learned to controll their leaks and not feed rumors.
1 Like
i was also thinking of changing out to this card, but after reading the post CES articles, and i saw somewhere, drawing a blank on the where, but supposedly AMD is only make like 5000 of these cards, and not seeding any to the AIBs. so with all that info, this card seeming like an, unfortunate, stop-gap till navi is ready which at this point many are hoping AMD actually announces at siggraph.
AMD has already released a statement to discredit the claims that they will only make 5000 and not give to AIBs. The AMD statement is as follows,
“While we don’t report on production numbers externally,” the statement reads, “we will have products available via AIB partners and AMD.com at launch of Feb. 7, and we expect Radeon VII supply to meet demand from gamers.”
Just wait till we get reviews from @wendell and other tech influencers.
2 Likes
Well that’s interesting… and goes directly against what they said at CES. AMD staffers said partner boards were unlikely… AIB partners at CES also confirmed that at CES they didn’t think they were going to get any.
ya, the whole thing is so murky, the only reasonable option is to wait for the dust to settle
I also heard the 5000 number but rather took it that it was 5000 currently ready, not 5000 total. though I did head the no AIB thing too, all third part from Jayz.
1 Like
I’m curious if they left SR-IOV enabled in the Radeon VII. As the Instinct card this is based on has it.
Not expecting it, but there’s some hope at least… https://twitter.com/RyanSmithAT/status/1084174168835289088
3 Likes
SR-IOV would be awesome, but if this card lacks the reset bug I’ll probably buy it (and sell my Vega 64).
A bit off topic but would SR-IOV allow you to run multiple VMs with the same card passed through for display output? I was googling around the other day but I couldn’t find any information if the libvirt/QEMU/KVM stack would support such a thing.
Yes, SR-IOV is exactly for one card many VMs situations. I don’t know if the current virtualising software takes advantage of it right now but if the hardware for it got into the publics hands and not just the 4 to 5 figure price range Enterprise GPUs, you would see it supported in very short order.
If this has SR-IOV enabled it would mean bypassing the reinitialization bug.
But am i right to asume the reinitialization bug was fixed on the nitro+ vega?
1 Like
I’ve got a Sapphire Nitro+ and the card (and VM) will lock up if I reboot the VM.
I also have a problem where UEFI initializes it at boot too, and I have to reset it via the sleep method before trying to run the VM. Otherwise it just hangs the card and the VM.
1 Like
I’m very tempted as I can then do VFIO with my Vega 64 with 2 different cards and retire the rx480 I have for that purpose.
Also as above it may fix the GPU reset bug. And if it has sr-iOV … gravy.
Ty, this will affect my choice of gpu purchase this spring.
just 3 dayes left, But with only performance testing being held back. Has anyone heard anything about sr-iov?
Not yet, but if SR-IOV is a thing this is a 100% buy from me.
edit:
I’ve logged a support ticket with AMD for information on Radeon VII SR-IOV. I’ll let you all know when i get a response.
3 Likes
First performance video is out, This one by Linus: