the 390 is just the 290x rebrand. The real question is the Radeon Fury cards
AMD: Pirate Islands (R* 3** series) Speculation/Rumor Thread
- Thread starter iMacmatician
- Start date
-
- Tags
- amd
Hope that the 290X rebrand will ship with 8GB?
(Edit: that's 3 similar replies in 30s!)
So, if i follow that down to it's essence, this argument is based via 3 or so intermediate stops on AMDs driver certification for OpenCL 2.0:
http://developer.amd.com/tools-and-...sdk/system-requirements-driver-compatibility/
Which basically only means what they bother to certify for OpenCL and what not.
D
Deleted member 2197
Guest
I think it was recently reviewed here ....
http://www.overclock3d.net/reviews/gpu_displays/xfx_r9_290x_8gb_dd_review/1
http://www.overclock3d.net/reviews/gpu_displays/xfx_r9_290x_8gb_dd_review/1
gamervivek
Regular
GCN1.1 and further is only said to support OpenCl2.0 by AMD. And they seem to have introduced it first in those laptop chips if Bonaire wasn't the first and it went unnoticed.
The second is that M370X beats a rather higher clocked(800 to 925) same specced M275 silly and is within 20% of a 950M at about same power envelope. And notebookcheck mention,
http://www.notebookcheck.net/AMD-Radeon-R9-M370X.142763.0.html
AMD also swore that Sea Islands wasn't an architecture name but included things like Oland, and snuck in TrueAudio into the Bonaire 7790 (and I believe it's not active on those products). Their answers to certain types of architectural questions should not be considered absolute, especially if for some reason they wanted to hide anything.
Meanwhile registers and product IDs don't lie.
D
Deleted member 13524
Guest
I just took our review sample HD 5770, so it is basically a 1st or 2nd-batch Juniper part (first batch went to devs seeding DX11) and installed the last WHQL (Catalyst Omega 14.12).
Turns out it does indeed report OpenCL 2.0 compliancy. So even with a moustache, I call the AMD-spec-page argument invalid.
Turns out it does indeed report OpenCL 2.0 compliancy. So even with a moustache, I call the AMD-spec-page argument invalid.
D
Deleted member 13524
Guest
Maybe they also upgraded the entire Evergreen family 4 years ago without anyone noticing.
I stand by my earlier assumption: AMD's latest drivers bring support for OpenCL 2.0 in any OpenCL 1.2-compliant device. It's just that OpenCL 2.0-specific functions run in the CPU and that's it.
I stand by my earlier assumption: AMD's latest drivers bring support for OpenCL 2.0 in any OpenCL 1.2-compliant device. It's just that OpenCL 2.0-specific functions run in the CPU and that's it.
I don't think that's possible. OCL doesn't abstract over compute devices. Hence only the software developer can decide which workload runs on which device
D
Deleted member 13524
Guest
And the driver can't simply redirect that workload to the CPU?
D
Deleted member 13524
Guest
Supposedly leaked R7 370 slide:
20% over Geforce 750 Ti (GM107). Sounds about right for Pitcairn:
They changed it from the R9 family to the R7 family, at least..
Also, is that a shorter PCB?
20% over Geforce 750 Ti (GM107). Sounds about right for Pitcairn:
They changed it from the R9 family to the R7 family, at least..
Also, is that a shorter PCB?
No, full pitcain, ie 270x is closer to 50% faster than 750ti. http://www.computerbase.de/2015-05/17-grafikkarten-von-amd-und-nvidia-im-vergleich/2/
Could be a salvage version of course.
Alessio1989
Regular
Pitcairn strikes back! 
FWIW, in our performance Index, R9 270 is 25% faster than 750 Ti with a small part of it due to larger framebuffer which helps in 4k. In the unpublished, isolated 1080p-index R9 270 is 23 percent faster thatn 750 Ti. Hopes are dwindling.
D
Deleted member 13524
Guest
I thought GCN GPUs compared favorably to Maxwell in GTA V, so I assumed those 20% advantage to be a "best-case" scenario.
