Some actual details on GFX1010 aka NAVI was just revealed by submits to LLVM.
github.com
github.com
For those of you wondering: It's another iteration of GCN, essentially improved VEGA with some new features. It's not some ground-up new architecture. As for launch-date, it's probably 3-4 months away given AMDs history of time between submits to LLVM/Mesa/kernel and GPU releases.
I read that as gtx 1010 and thought "what retardedass stupid shit is happening this time"
I was disappointed honestly
sony made navi.
it will be tweaked for gaymen.
>it's GCN
Whoever actually thought it was going to be a new arch is fucking braindead.
Hey I got a scoop on Navi it's going to be utter shit just like every other AMD gpu for the past 10 years wow isn't that amazing
>3-4 months away
AMDs referred to their GPUs as GFX[X](Y). The AMD Polaris cards like 480, 580 and 590 are GFX8 cards in all the source-code. It's not used in marketing or anything like that, it's only useful if you look at the actual drivers. If you want to know what registers NAVIs you'll be looking for GFX1010.
HBM MXM when?
7970, 280 and 380 are good cards.
navi is not vega because they dropped vega and raja.
no hbm no buy
>def FeatureLdsMisalignedBug : SubtargetFeature"Some GFX10 bug with misaligned multi-dword LDS access in WGP mode"
That's gold, very impressed by this
>card not close to release
>already working around hardware bugs in LLVMs code
>navi is not vega
From what I can that's exactly what NAVI is. Look at all the checks that were changed from isGFX9 (VEGA) to isGFX9() || isGFX10() (VEGA or NAVI).
That's their next line of professional cards based on CGN / Vega, since Navi won't get a professional release right away, being a new arch.
Plus Navi won't replace all CGN cards.
>Plus Navi won't replace all CGN cards.
Instantly that is. Eventually, sure.
Should I pick up a GCN 4.0 over my shitty ass Tonga MXM?
It's like 300 bucks but I'm having issues with EDID on the current card and the free and open drivers seem to disregard Tonga.
>It's another iteration of GCN, essentially improved VEGA with some new features. It's not some ground-up new architecture.
We've known this forever, dumbass. It's not some hot scoop.
>It's like 300 bucks
I guess it depends on what kind of laptop you need a MXM card for - but it seems like an expensive upgrade. Again, depends on the machine.
>it's probably 3-4 months away given AMDs history of time between submits to LLVM/Mesa/kernel and GPU releases.
inb4 lack of mention during computex, amdoredtv caught telling lies again and another 1/2 year delay
amd cfo basicly said their new uarch will come in 2020-21
>It's another iteration of GCN
Who could've guessed?
Looking forward to it. As long as NVIDIA can't run two monitors with different refresh rates at the same time without issues, they're not getting my money.
Is this still an issue? I do it all the time at work. We're constantly jumping around 29.97hz, 30, 59.94 and 60 in our main display and have the second at 60. Are you going above 60 as I haven't tried much there but the devices we use do sometimes auto set 120hz.
I see people complaining about this issue regularly on forums earlier this year. Not sure about the current state of things.
god this board is retarded
>As for launch-date, it's probably 3-4 months away given AMDs history of time between submits to LLVM/Mesa/kernel and GPU releases.
That would pretty much match up with the rumor that it's going to launch at Gamescon wich is middle/late August.
>tfw 4870 is more than 10 years old now
Loved that gpu
I, too, enjoy games of the video variety.
So VII will be the best AMD card?
Until Navi 20 next year it would seem.
Calling user error or its more likely the screens themselves. Using the nvidia control panel it's pretty easy if a bit slow, we're even using it with custom resolutions in the past. But these days with hdmi2 most screens support TV and monitor resolutions and the control panel exposes all of them. Haven't even attempted it in the cancer that is the nvexperience software. The limit we hit is using dsr and custom resolutions at the same time.
Radeon control panel is missing many good resolution options when I tried it, it hides too much.
You will enjoy em when you grow up.
Just like Vega is GCN but the SPUs are totally different?
Also:
>The Vega architectural whitepaper notes that Vega's next-generation geometry path can process more than 17 primitives per clock, compared to the four primitives per clock that the Vega 10 GPU can process using the traditional Direct3D rendering path.
It means that Vega full potential isnt even being used due to Driver/API constrains. Maybe that's why World War Z vulkan renderer makes it beat Nvidia.
The Radeon Software have timing edits for displays.
>As long as NVIDIA can't run two monitors with different refresh rates at the same time without issues
What issues? I regularly run one screen at 24Hz and the other 2 at 60Hz. I haven't noticed any difference compared to when I had AMD.
Yeah, AMD pretty much said so when they released Radeon VII too.
Don't forget leafblower 290x
Man I loved that card. Couldn't oc 50mhz but still kicked ass
Yeah, just bought a used 1080 ti because the wait was finally too much to avoid paying the evil team. Looks like it'll last a couple years yet before AMD can compete
>def FeatureNSAEncoding : SubtargetFeature "HasNSAEncoding",
> "true",
> "Support NSA encoding for image instructions"
> >;
Oh no, they overtook AMD.
If you are going to advertise like that at least get good looking models with nice racks. Those two are pitiful.
25 and single sexual partner detected
AMD really can't do anything right.