SemiAccurate Forums  

 
Go Back   SemiAccurate Forums > Search Forums

Showing results 1 to 25 of 311
Search took 0.01 seconds.
Search: Posts Made By: gruffi
Forum: GPUs 01-26-2014, 07:19 AM
Replies: 109
Views: 13,577
Posted By gruffi
Nvidia? No way. :D

Nvidia? No way. :D
Forum: CPUs 01-13-2014, 12:32 AM
Replies: 3,536
Views: 452,742
Posted By gruffi
IPC should be above Phenom 2 while still having a...

IPC should be above Phenom 2 while still having a 20-25% clock advantage.
Forum: CPUs 01-12-2014, 01:44 PM
Replies: 463
Views: 54,792
Posted By gruffi
Kaveri's L2 is quite large and fast. At least...

Kaveri's L2 is quite large and fast. At least much faster than Deneb's L3. A significant improvement over Trinity/Richland.
Forum: CPUs 01-12-2014, 01:22 PM
Replies: 3,536
Views: 452,742
Posted By gruffi
I don't think so. Cinebench R11.5 uses the same...

I don't think so. Cinebench R11.5 uses the same FPU but looks much better. ;) There seems to be something in Cinebench R15 that Steamroller doesn't like. Does ICC fool us once again?
Forum: CPUs 01-10-2014, 11:07 AM
Replies: 463
Views: 54,792
Posted By gruffi
Slow? Atom is slow, not Bulldozer. ;) ...

Slow? Atom is slow, not Bulldozer. ;)


There's no trace cache in Steamroller. There's only a loop buffer. Probably similar to what was implemented in Jaguar. AMD stated it's a power saving...
Forum: CPUs 01-09-2014, 10:38 AM
Replies: 3,536
Views: 452,742
Posted By gruffi
Probably he means "28nm SHP has a TDP vs Clock...

Probably he means "28nm SHP has a TDP vs Clock Rate issue compared to 32nm SOI". Which I wouldn't call an issue. The 28nm bulk process was just optimized in a different way, more APU focused. But it...
Forum: CPUs 01-09-2014, 10:05 AM
Replies: 463
Views: 54,792
Posted By gruffi
Actually it's not that bad. Kaveri's L2 should be...

Actually it's not that bad. Kaveri's L2 should be even better. Latency is only a problem if you cannot sustain a high pipeline utilization. I think that's not a problem for Bulldozer.


Core 2...
Forum: CPUs 01-08-2014, 04:09 PM
Replies: 463
Views: 54,792
Posted By gruffi
Not that fast. Bulldozer's L2 read throughput is...

Not that fast. Bulldozer's L2 read throughput is actually higher (http://www.xtremesystems.org/forums/showthread.php?286212-new-AIDA-64-seems-fixed-for-FX-CPUs) than Intel's. Intel has an advantage...
Forum: CPUs 01-07-2014, 06:15 PM
Replies: 990
Views: 224,893
Posted By gruffi
You actually said: How can you increase...

You actually said:

How can you increase performance per watt with dual channel if the processor doesn't support it? That won't happen. Performance per watt will be increased due to other...
Forum: CPUs 01-07-2014, 11:50 AM
Replies: 990
Views: 224,893
Posted By gruffi
I just wanted to make clear that Kabini or Beema...

I just wanted to make clear that Kabini or Beema doesn't support dual channel. Even if the board does. If you want to speculate about future APUs with dual channel support I guess it's the wrong...
Forum: CPUs 01-07-2014, 11:41 AM
Replies: 463
Views: 54,792
Posted By gruffi
Do they? Intel needs L3 cache because they have...

Do they? Intel needs L3 cache because they have extremely small L2 caches. In my opinion the cache of Core Duo and Core 2 Duo was more effective than Intel's cache today. Actually it's quite funny...
Forum: GPUs 01-07-2014, 11:02 AM
Replies: 109
Views: 13,577
Posted By gruffi
Who says that most monitors don't have it yet?...

Who says that most monitors don't have it yet? Yes, it needs monitor support. But that is not the point. The point is, it's a standardized VESA feature. That makes it much easier to support.

...
Forum: CPUs 01-06-2014, 06:26 PM
Replies: 463
Views: 54,792
Posted By gruffi
I don't think that AMD will ever consider an L3...

I don't think that AMD will ever consider an L3 cache for APUs. At least not for client APUs. L3 cache is a server thing. Their current cache hierarchy is fine. The Jaguar CU even shows that they can...
Forum: CPUs 01-06-2014, 06:14 PM
Replies: 990
Views: 224,893
Posted By gruffi
No. I just asked my crystal ball. :rolleyes: ...

No. I just asked my crystal ball. :rolleyes:

Man, some people are really impolite. You should follow the discussion and don't react like a little kid. I was referring to "dual channel" that Caza...
Forum: CPUs 01-06-2014, 03:03 PM
Replies: 990
Views: 224,893
Posted By gruffi
These boards support Kaveri, Richland and...

These boards support Kaveri, Richland and Trinity. No Kabini or Beema.
Forum: CPUs 01-06-2014, 02:39 PM
Replies: 463
Views: 54,792
Posted By gruffi
But the changes of the CPU architecture might be...

But the changes of the CPU architecture might be even bigger on Excavator. 256-bit FMACs, wider integer clusters (2-way SMT?), AVX2, etc.
Forum: CPUs 01-04-2014, 11:28 AM
Replies: 3,536
Views: 452,742
Posted By gruffi
That's not the task of Sandra. Sandra uses OpenCL...

That's not the task of Sandra. Sandra uses OpenCL or whatever. It doesn't need to generate HSAIL code to support HSA. HSAIL is the layer between high level languages and hardware ISA.



Kaveri's...
Forum: CPUs 01-04-2014, 11:07 AM
Replies: 463
Views: 54,792
Posted By gruffi
And it supports AVX2.

And it supports AVX2.
Forum: CPUs 01-03-2014, 08:11 PM
Replies: 3,536
Views: 452,742
Posted By gruffi
Who says there's no HSAIL code executed?

Who says there's no HSAIL code executed?
Forum: CPUs 01-03-2014, 07:31 PM
Replies: 3,536
Views: 452,742
Posted By gruffi
Using the iGPU means HSA. ;)

Using the iGPU means HSA. ;)
Forum: CPUs 01-03-2014, 05:47 PM
Replies: 3,536
Views: 452,742
Posted By gruffi
BUT that seems to be the power consumption of...

BUT that seems to be the power consumption of 6800K at 3.5 GHz, not 4.1-4.4 GHz. ;) And it's still 5W (load-idle) higher than Kaveri. 7850K with 200-500 MHz more definitely shouldn't need as much...
Forum: CPUs 01-02-2014, 02:27 PM
Replies: 3,536
Views: 452,742
Posted By gruffi
Not false. There are still no concrete...

Not false. There are still no concrete frequencies for CPU and GPU mentioned.


Which are still speculations and nothing that AMD announced.


Just the old stuff repeated again and again. Who...
Forum: CPUs 01-02-2014, 12:59 PM
Replies: 463
Views: 54,792
Posted By gruffi
@nvidiot You still shredder this thread with...

@nvidiot

You still shredder this thread with pure nonsense. Carrizo is not the competition of KNL, Broadwell EP or Kepler. Carrizo is a client design for desktop and mobile systems. pTmd is right....
Forum: CPUs 01-02-2014, 11:21 AM
Replies: 3,536
Views: 452,742
Posted By gruffi
No, they didn't. First, I don't see any concrete...

No, they didn't. First, I don't see any concrete numbers on the slide. Second, that's just related to the architecture pipeline. A real chip is another story because of its manufacturing process. ;)...
Forum: CPUs 01-02-2014, 10:51 AM
Replies: 463
Views: 54,792
Posted By gruffi
Which weak APUs? AMD has the fastest APUs on the...

Which weak APUs? AMD has the fastest APUs on the market. For everything above they still have their dGPUs. Which will fit flawlessly in the HSA infrastructure. If you didn't know. ;)

KNL and...
Showing results 1 to 25 of 311

 
Forum Jump

All times are GMT -5. The time now is 11:07 AM.


Powered by vBulletin® Version 3.8.7
Copyright ©2000 - 2014, vBulletin Solutions, Inc.
SemiAccurate is a division of Stone Arch Networking Services, Inc. Copyright 2009 Stone Arch Networking Services, Inc, all rights reserved.