AMD CPU speculation... and expert conjecture

[Reepca]

Apologies if it's already been stated earlier in the thread (it's a lot to read through), but does anyone know if AMD plans to continue production of their FX chips (that is, desktop non-APU CPUs)? Or if it's even a possibility? Steamroller, if I understood correctly, gained a respectable IPC increase over Piledriver, so is there a chance that AMD could produce some Steamroller CPUs? Or Excavator? I understand that HSA is certainly more efficient, but the point remains that you can only fit so much into one chip - as much as I would love to have an AMD Steamroller 8-CPU-core-290X-level-GPU APU, it isn't happening anytime soon.

On another note, does anyone know how much mantle affects Dual Graphics performance? The reason Dual Graphics wouldn't normally work very well is that, by design, the APU should have a balance of CPU and GPU so that neither of them bottleneck the other. Doubling the net GPU computing power would cause a bottleneck. But where Mantle, as far as I have read, shines the most is where the CPU is the limiting factor - which sounds like Dual Graphics to me!

 

[Cazalan]

Just remember: Fusion is ALWAYS 20 years away.

Yeah because our leaders are idiots. A trillion dollar war to keep some control in oil lands. If a fraction of that was put into fusion research it would have been solved already.

 

[jdwii]

Apologies if it's already been stated earlier in the thread (it's a lot to read through), but does anyone know if AMD plans to continue production of their FX chips (that is, desktop non-APU CPUs)? Or if it's even a possibility? Steamroller, if I understood correctly, gained a respectable IPC increase over Piledriver, so is there a chance that AMD could produce some Steamroller CPUs? Or Excavator? I understand that HSA is certainly more efficient, but the point remains that you can only fit so much into one chip - as much as I would love to have an AMD Steamroller 8-CPU-core-290X-level-GPU APU, it isn't happening anytime soon.

On another note, does anyone know how much mantle affects Dual Graphics performance? The reason Dual Graphics wouldn't normally work very well is that, by design, the APU should have a balance of CPU and GPU so that neither of them bottleneck the other. Doubling the net GPU computing power would cause a bottleneck. But where Mantle, as far as I have read, shines the most is where the CPU is the limiting factor - which sounds like Dual Graphics to me!

Ok the FX series will be continuing but not on the AM3+ platform i'm not sure if they want to continue using that name for their new generation CPU's coming out in 2016(maybe 17 Amd delays things a lot). I'm pretty sure we won't be seeing any updates to their current architecture on the AM3+ platform. Rumors go around some times and speak about a possible 6 core APU from them using Steamroller or excavator but i'm not to sure if that will ever happen.

I'm sure you already knew that Amd is advertising their GPU's more then anything else and using HSA to support that. Its still isn't important yet to normal consumers or gamers but that might change in the future. Also i expect Mantle to improve crossfire performance over it lowing the CPU bottleneck.

 

[Reepca]

Apologies if it's already been stated earlier in the thread (it's a lot to read through), but does anyone know if AMD plans to continue production of their FX chips (that is, desktop non-APU CPUs)? Or if it's even a possibility? Steamroller, if I understood correctly, gained a respectable IPC increase over Piledriver, so is there a chance that AMD could produce some Steamroller CPUs? Or Excavator? I understand that HSA is certainly more efficient, but the point remains that you can only fit so much into one chip - as much as I would love to have an AMD Steamroller 8-CPU-core-290X-level-GPU APU, it isn't happening anytime soon.

On another note, does anyone know how much mantle affects Dual Graphics performance? The reason Dual Graphics wouldn't normally work very well is that, by design, the APU should have a balance of CPU and GPU so that neither of them bottleneck the other. Doubling the net GPU computing power would cause a bottleneck. But where Mantle, as far as I have read, shines the most is where the CPU is the limiting factor - which sounds like Dual Graphics to me!

some games yes some no
http://www.hardcoreware.net/radeon-dual-graphics-performance-kaveri/4/

I see. But the line (last page) that I noticed the most was "Seems to be a "best case scenario" for Dual Graphics, but Mantle doesn't work yet.", describing Battlefield 4 at the time. I suppose the ideal test would be to test Dual Graphics on BF4 without Mantle and then with Mantle. Are there any other, more recent statistics I could look at (that you know of)?

Additionally, I imagine that with the newer drivers available now, perhaps we will see more gains in other games?

 

[blackkstar]

@blackstar actually CMT is one of few things which works as it should in Bulldozer. People were looking for cause of poor performance, so they took the most hyped feature. And no one cares that BD has for example smaller L1, bigger L2, less ALUs/AGUs than K10.

Yes that is often overlooked it really comes down to what your saying.
I honestly upgraded to the 8350 to see for myself and its L3 cache is one major issue and the L2 cache one programmer even claimed its easier to just treat the CPU as a 4 core since 1 core can steal all the L2 cache in 1 module.

Also as stated by palladin9479(and many others) the scheduler/branch predictor is also an issue. Hopefully Amd does what Intel did and includes a µop cache(maybe decrease their pipeline length to). Furthermore i hope they stop sharing the L1 cache.

CMT will still lower single threaded performance its just impossible to claim otherwise when it shares resources performance will degrade even if its by 5% one day(20% now 30% in some cases). The only way to overcome that is by brute forcing your design like adding lots of cache or increasing frequency by a decent amount. Edit actually if the software you use can take advantage of sharing resources it might actually overcome the 20% hit and perhaps boost performance.

CMT also introduces latency issues and overall it might just be easier for Amd or others to use SMT since its free to add on and pushes efficiency and won't make single threaded performance worse(if the OS/software handles it right).
I see CMT being more useful for many-core designs where single threaded performance doesn't matter as much like servers.

I suppose in my round-about way, I was saying that CMT is the basic principle of sharing core resources among other cores to reduce die size, power consumption, cost to manufacture, etc.

The problem with the way AMD does things is that they sacrificed a lot of what you're talking about in order to reduce die size and stuff to increase core count.

The new Oracle design has some sharing going on, but it looks more balanced, judging by slides, and unless I'm not understanding things, you can get one thread per module. meaning all those resources can be used by a single thread. So with Oracle CMT, you end up with a single thread using more resources than it normally would have. And with AMD CMT you end up with one thread with maybe having enough resources left over.

 

[gamerk316]

Apologies if it's already been stated earlier in the thread (it's a lot to read through), but does anyone know if AMD plans to continue production of their FX chips (that is, desktop non-APU CPUs)? Or if it's even a possibility? Steamroller, if I understood correctly, gained a respectable IPC increase over Piledriver, so is there a chance that AMD could produce some Steamroller CPUs? Or Excavator? I understand that HSA is certainly more efficient, but the point remains that you can only fit so much into one chip - as much as I would love to have an AMD Steamroller 8-CPU-core-290X-level-GPU APU, it isn't happening anytime soon.

On another note, does anyone know how much mantle affects Dual Graphics performance? The reason Dual Graphics wouldn't normally work very well is that, by design, the APU should have a balance of CPU and GPU so that neither of them bottleneck the other. Doubling the net GPU computing power would cause a bottleneck. But where Mantle, as far as I have read, shines the most is where the CPU is the limiting factor - which sounds like Dual Graphics to me!

some games yes some no
http://www.hardcoreware.net/radeon-dual-graphics-performance-kaveri/4/

I see. But the line (last page) that I noticed the most was "Seems to be a "best case scenario" for Dual Graphics, but Mantle doesn't work yet.", describing Battlefield 4 at the time. I suppose the ideal test would be to test Dual Graphics on BF4 without Mantle and then with Mantle. Are there any other, more recent statistics I could look at (that you know of)?

Additionally, I imagine that with the newer drivers available now, perhaps we will see more gains in other games?

Dual graphics, from an HSA perspective, really doesn't make any sense. You either have the APU being starved due to the memory bandwidth bottleneck, or the discrete GPU being starved due to the latency involved in copying data across the PCI-E bus. That's why you don't see the same scaling you do with SLI or CF.

Mantle does NOTHING to address this issue. Mantle is simply a way of reducing the CPU bottleneck in games, so when in a situation where the GPU has headroom, you get an increase in FPS based on the added performance the CPU has to perform.

 

[szatkus]

Apologies if it's already been stated earlier in the thread (it's a lot to read through), but does anyone know if AMD plans to continue production of their FX chips (that is, desktop non-APU CPUs)? Or if it's even a possibility? Steamroller, if I understood correctly, gained a respectable IPC increase over Piledriver, so is there a chance that AMD could produce some Steamroller CPUs? Or Excavator? I understand that HSA is certainly more efficient, but the point remains that you can only fit so much into one chip - as much as I would love to have an AMD Steamroller 8-CPU-core-290X-level-GPU APU, it isn't happening anytime soon.

On another note, does anyone know how much mantle affects Dual Graphics performance? The reason Dual Graphics wouldn't normally work very well is that, by design, the APU should have a balance of CPU and GPU so that neither of them bottleneck the other. Doubling the net GPU computing power would cause a bottleneck. But where Mantle, as far as I have read, shines the most is where the CPU is the limiting factor - which sounds like Dual Graphics to me!

some games yes some no
http://www.hardcoreware.net/radeon-dual-graphics-performance-kaveri/4/

I see. But the line (last page) that I noticed the most was "Seems to be a "best case scenario" for Dual Graphics, but Mantle doesn't work yet.", describing Battlefield 4 at the time. I suppose the ideal test would be to test Dual Graphics on BF4 without Mantle and then with Mantle. Are there any other, more recent statistics I could look at (that you know of)?

Additionally, I imagine that with the newer drivers available now, perhaps we will see more gains in other games?

Dual graphics, from an HSA perspective, really doesn't make any sense. You either have the APU being starved due to the memory bandwidth bottleneck, or the discrete GPU being starved due to the latency involved in copying data across the PCI-E bus. That's why you don't see the same scaling you do with SLI or CF.

Mantle does NOTHING to address this issue. Mantle is simply a way of reducing the CPU bottleneck in games, so when in a situation where the GPU has headroom, you get an increase in FPS based on the added performance the CPU has to perform.

Not nothing.

http://www.rebellion.co.uk/blog/2014/10/2/mantle-comes-to-sniper-elite-3

Read "Conclusions"

 

[gamerk316]

Apologies if it's already been stated earlier in the thread (it's a lot to read through), but does anyone know if AMD plans to continue production of their FX chips (that is, desktop non-APU CPUs)? Or if it's even a possibility? Steamroller, if I understood correctly, gained a respectable IPC increase over Piledriver, so is there a chance that AMD could produce some Steamroller CPUs? Or Excavator? I understand that HSA is certainly more efficient, but the point remains that you can only fit so much into one chip - as much as I would love to have an AMD Steamroller 8-CPU-core-290X-level-GPU APU, it isn't happening anytime soon.

On another note, does anyone know how much mantle affects Dual Graphics performance? The reason Dual Graphics wouldn't normally work very well is that, by design, the APU should have a balance of CPU and GPU so that neither of them bottleneck the other. Doubling the net GPU computing power would cause a bottleneck. But where Mantle, as far as I have read, shines the most is where the CPU is the limiting factor - which sounds like Dual Graphics to me!

some games yes some no
http://www.hardcoreware.net/radeon-dual-graphics-performance-kaveri/4/

I see. But the line (last page) that I noticed the most was "Seems to be a "best case scenario" for Dual Graphics, but Mantle doesn't work yet.", describing Battlefield 4 at the time. I suppose the ideal test would be to test Dual Graphics on BF4 without Mantle and then with Mantle. Are there any other, more recent statistics I could look at (that you know of)?

Additionally, I imagine that with the newer drivers available now, perhaps we will see more gains in other games?

Dual graphics, from an HSA perspective, really doesn't make any sense. You either have the APU being starved due to the memory bandwidth bottleneck, or the discrete GPU being starved due to the latency involved in copying data across the PCI-E bus. That's why you don't see the same scaling you do with SLI or CF.

Mantle does NOTHING to address this issue. Mantle is simply a way of reducing the CPU bottleneck in games, so when in a situation where the GPU has headroom, you get an increase in FPS based on the added performance the CPU has to perform.

Not nothing.

http://www.rebellion.co.uk/blog/2014/10/2/mantle-comes-to-sniper-elite-3

Read "Conclusions"

Which I disagree with. Look at the low res results: with the R9 290x: 19.1 and 19.2 percent gain via Mantle. That's your reduced CPU overhead in non-GPU bottlenecked situations. The Q9450, by contrast, appears to hit a CPU bottleneck again. But that's the extra CPU power coming into play, not dual graphics.

 

[colinp]

That article from Szatkus' link was very interesting. It's clear that Mantle is a game changer in that it is paving the ways for other APIs. And in itself is delivering benefits in terms of CPU utilisation and image quality.

And the speculation about how Mantle can open up possibilities for dual graphics too (in the conclusion that Gamerk ignored because it refuted his earlier claim about Mantle doing "NOTHING" for dual graphics).

I for one, welcome disruptors to any industry, as otherwise you get left with DX11, Internet Explorer 6 and the Start Menu with nothing ever changing. (Ok, forget that last one.)

 

[noob2222]

Thats part of the problem with CMT, it was designed for servers first and DT 2nd. Thats where the penalty for using both cores in a module offset in specific usage. In a server application, the penalty isnt as detrimental as it is for gaming. Thats also why the windows scheduler was updated for CMT to utilize 4 cores before modules.

The problem with it is people would rather look at worst case scenarios, making the CMT approach look worse as a DT oreinted processor.

 

[gamerk316]

Thats also why the windows scheduler was updated for CMT to utilize 4 cores before modules.

The irony again being if AMD simply set the HTT bit in the CPUID, this wouldn't have even been needed. Seemed kind of silly to wait for MSFT to patch the problem when AMD could have solved it themselves. All the scheduler patch does is treat BD based processors the same way Intel's HTT enabled processors are (Use the HTT/CMT cores last).

 

[Reepca]

Apologies if it's already been stated earlier in the thread (it's a lot to read through), but does anyone know if AMD plans to continue production of their FX chips (that is, desktop non-APU CPUs)? Or if it's even a possibility? Steamroller, if I understood correctly, gained a respectable IPC increase over Piledriver, so is there a chance that AMD could produce some Steamroller CPUs? Or Excavator? I understand that HSA is certainly more efficient, but the point remains that you can only fit so much into one chip - as much as I would love to have an AMD Steamroller 8-CPU-core-290X-level-GPU APU, it isn't happening anytime soon.

On another note, does anyone know how much mantle affects Dual Graphics performance? The reason Dual Graphics wouldn't normally work very well is that, by design, the APU should have a balance of CPU and GPU so that neither of them bottleneck the other. Doubling the net GPU computing power would cause a bottleneck. But where Mantle, as far as I have read, shines the most is where the CPU is the limiting factor - which sounds like Dual Graphics to me!

some games yes some no
http://www.hardcoreware.net/radeon-dual-graphics-performance-kaveri/4/

I see. But the line (last page) that I noticed the most was "Seems to be a "best case scenario" for Dual Graphics, but Mantle doesn't work yet.", describing Battlefield 4 at the time. I suppose the ideal test would be to test Dual Graphics on BF4 without Mantle and then with Mantle. Are there any other, more recent statistics I could look at (that you know of)?

Additionally, I imagine that with the newer drivers available now, perhaps we will see more gains in other games?

Dual graphics, from an HSA perspective, really doesn't make any sense. You either have the APU being starved due to the memory bandwidth bottleneck, or the discrete GPU being starved due to the latency involved in copying data across the PCI-E bus. That's why you don't see the same scaling you do with SLI or CF.

Mantle does NOTHING to address this issue. Mantle is simply a way of reducing the CPU bottleneck in games, so when in a situation where the GPU has headroom, you get an increase in FPS based on the added performance the CPU has to perform.

I suppose that my reasoning was that (looking at this somewhat abstractly) having a balanced system in terms of GPU vs CPU capabilities is essential to maximizing performance, and that by design an APU alone should be balanced, and therefore that making the GPU effectively stronger (adding in a discrete GPU) would cause a bottleneck due to the CPU being too weak compared to the combined performance of the integrated and discrete GPU. Since, as you say, "Mantle is simply a way of reducing the CPU bottleneck in games", wouldn't it (in theory) have a larger impact (% uplift) on systems using Dual Graphics, since (as far as I understand) by the principle of it the system should be CPU bottlenecked?

But if I understand you correctly, any bottleneck alleviated wouldn't make any difference due to less-abstract issues like memory bandwidth and latency preventing the GPU capability out-growing the CPU?

Also, thanks to the guy who posted that link to the Rebellion article, very interesting to read!

 

[gamerk316]

I suppose that my reasoning was that (looking at this somewhat abstractly) having a balanced system in terms of GPU vs CPU capabilities is essential to maximizing performance, and that by design an APU alone should be balanced, and therefore that making the GPU effectively stronger (adding in a discrete GPU) would cause a bottleneck due to the CPU being too weak compared to the combined performance of the integrated and discrete GPU. Since, as you say, "Mantle is simply a way of reducing the CPU bottleneck in games", wouldn't it (in theory) have a larger impact (% uplift) on systems using Dual Graphics, since (as far as I understand) by the principle of it the system should be CPU bottlenecked?

But if I understand you correctly, any bottleneck alleviated wouldn't make any difference due to less-abstract issues like memory bandwidth and latency preventing the GPU capability out-growing the CPU?

Also, thanks to the guy who posted that link to the Rebellion article, very interesting to read!

The issue is this: The AMD APU is being starved not because of CPU performance, but because of the latency involved in copying data across the memory bus. No amount of CPU improvements will solve that bottleneck. As a result, in any case where the AMD APU is NOT CPU bound, any CPU improvements will be expected to give linear improvements in performance; EG, a 5% CPU boost will yield about 5% performance. That's also why higher tier chips don't gain as much via Mantle. Since you have a GPU bottleneck, the improved CPU performance also tends to yield linear increases.

If however you currently have a CPU bottleneck, and Mantle pushes performance to the point where the bottleneck moves to the GPU instead, you'll see performance increases greater then the CPU side benefit Mantle brings. That's the case where improvement percentages can reach off the charts.

As for the "balanced system" argument, i always advocated going "big" on the CPU, then forgetting about it for the next 4-5 years or so, as that almost always ends up being cheaper then constantly upgrading. Hence my 2600k, which is still going quite strong. Hoping to get a few more years out of it; not like its an obsolete chip performance wise yet.

 

[jdwii]

Looks like Evil within is CPU bound and uses 2 cores and the game doesn't really care about having more then that.
http://gamegpu.ru/action-/-fps-/-tps/the-evil-within-test-gpu.html

Playing a game at 30fps i feel like i just bought my Nintendo Wii back in 2006 again.

 

[etayorius]

You seriously can`t expect much from any Beth title.

 

[Reepca]

The issue is this: The AMD APU is being starved not because of CPU performance, but because of the latency involved in copying data across the memory bus. No amount of CPU improvements will solve that bottleneck. As a result, in any case where the AMD APU is NOT CPU bound, any CPU improvements will be expected to give linear improvements in performance; EG, a 5% CPU boost will yield about 5% performance. That's also why higher tier chips don't gain as much via Mantle. Since you have a GPU bottleneck, the improved CPU performance also tends to yield linear increases.

If however you currently have a CPU bottleneck, and Mantle pushes performance to the point where the bottleneck moves to the GPU instead, you'll see performance increases greater then the CPU side benefit Mantle brings. That's the case where improvement percentages can reach off the charts.

As for the "balanced system" argument, i always advocated going "big" on the CPU, then forgetting about it for the next 4-5 years or so, as that almost always ends up being cheaper then constantly upgrading. Hence my 2600k, which is still going quite strong. Hoping to get a few more years out of it; not like its an obsolete chip performance wise yet.

In the immortal words of John Watson, "I still don't understand". If the APU by itself isn't bottlenecked either way (CPU or GPU), then adding a discrete GPU should make it CPU bottlenecked, right? Then wouldn't you "see performance increases greater than the CPU side benefit Mantle brings"? The AMD APU might not be starved because of CPU performance by itself, but when you add a discrete card wouldn't it make sense for it to be?

I suppose perhaps the problem is that I'm looking at this in terms of "CPU" and "GPU", but not including the other player - RAM, since in non-APU systems it's more or less a non-factor in terms of CPU and GPU performance. What I don't understand is how post-Kaveri it matters - if the CPU and GPU portions of the APU can use the exact same memory, what copying is there to be done? And if memory is the bottleneck, why doesn't fast RAM (2400+) yield any notable benefit?

 

[palladin9479]

The issue is this: The AMD APU is being starved not because of CPU performance, but because of the latency involved in copying data across the memory bus.

This isn't true of APU's. Copying the contents of one memory cell to another is ridiculously fast and GDDR5 has much greater latency then DDR3. Raw bandwidth is the killer for APU's. Each generation of AMD's design required more and more bandwidth to hit it's wall, which is actually one of the biggest issues with the 7850K vs everything else. There isn't enough bandwidth to keep 512 "GCN cores" occupied, 384 is right at the max to what you can economically feed. This might change with DDR4, though I don't expect much for at least another year or two. Kavari is basically a DDR3 R5 240/250 with two x86 modules bolted onto the side of it.

 

[Cazalan]

The Tonga style texture compression should help out with the limited RAM bandwidth.

AMD stock fell off a cliff. I just hope they can survive to 2016 to actually put out these new cores.

 

[Reepca]

http://www.corsair.com/en-us/blog/2014/february/understanding-kaveri

If I read this correctly, the reason that performance isn't scaling as well with faster RAM as it should is that AMD's memory controller struggles with translating faster RAM into more bandwidth. Am I right in stating that?

So with DDR4, would we see a new/improved memory controller? Or maybe with Carrizo/Excavator? It seems like the memory controller is what's holding back the GPU side's potential. Would that require a new socket or chipset?

 

[palladin9479]

He needed to overclock the CPU_NB, otherwise your going to get poor performance with AIDA64. Don't ask me why but AMD has the CPU_NB running at a much lower clock speed then it's capable of, and the IMC and prefetch system run inside there. My fx8350 has it's NB clocked at 2.6Ghz and gets numbers much higher then what he posted, same with my 6800K (2.2 or 2.4Ghz can't remember). Kaveri comes with the CPU_NB clocked at something piss poor 1.8, pretty much every OC guide recommends clocking it to 2.0 ~ 2.1Ghz. This was the only weakness I could find in the 7600, it doesn't let you change the CPU_NB.

I remember running some AIDA tests on my different chips and the CPU_NB speed had a very large effect on the results.

 

[Reepca]

So if one paired fast RAM (2400 speed, for example) with an overclocked CPU_NB, one could in theory expect better performance in real-world situations (than shown in that article... which was still quite a lot, over 40% increase in FPS from 1333 to 2640) as well - assuming that, for example, the a10-7850k was indeed badly memory-bandwidth-starved?

Neat. One of the considerations I was... well... considering regarding a potential upgrade to the 7850k was if it would be worth it to get anything above 2133 speed RAM, but would I be right in saying that a CPU_NB overclock should help the APU to scale well with the faster RAM?

 

[palladin9479]

The 7850K is horribly priced for it's performance level. The 7700K would be a better purchase, then pair it with good quality 2133 memory. You can get a 7850K with 2400 memory but the costs would exceed an 860K with a 1GB GDDR5 dGPU.

http://www.newegg.com/Product/Product.aspx?Item=N82E16814121800&cm_re=Radeon_250-_-14-121-800-_-Product
http://www.newegg.com/Product/Product.aspx?Item=N82E16814150711&cm_re=Radeon_250-_-14-150-711-_-Product
http://www.newegg.com/Product/Product.aspx?Item=N82E16814150676&cm_re=Radeon_250-_-14-150-676-_-Product

http://www.newegg.com/Product/Product.aspx?Item=N82E16819113379&cm_re=860K-_-19-113-379-_-Product

The existence of 80~100 USD GDDR5 dGPU's and good 80~90 USD CPU's place a practical limit on APU's. It's why I still view the A8-7600 and A8-6800K as the most value orientated purchases. Pair them both with cheap two sticks of 2133 memory and you got a solid platform that will meet 90% of what people need done.

 

[colinp]

Intel has posted some impressive numbers, and prempted AMD by saying PC sales are up. Lisa Su could be in for a baptism of fire tomorrow.

 

[gamerk316]

The issue is this: The AMD APU is being starved not because of CPU performance, but because of the latency involved in copying data across the memory bus.

This isn't true of APU's. Copying the contents of one memory cell to another is ridiculously fast and GDDR5 has much greater latency then DDR3. Raw bandwidth is the killer for APU's. Each generation of AMD's design required more and more bandwidth to hit it's wall, which is actually one of the biggest issues with the 7850K vs everything else. There isn't enough bandwidth to keep 512 "GCN cores" occupied, 384 is right at the max to what you can economically feed. This might change with DDR4, though I don't expect much for at least another year or two. Kavari is basically a DDR3 R5 240/250 with two x86 modules bolted onto the side of it.

That's what I meant. No idea why I put Latency instead of Bandwidth. Point remains the same though: The APU is being starved of data due to RAM.

 

[jdwii]

Intel has posted some impressive numbers, and prempted AMD by saying PC sales are up. Lisa Su could be in for a baptism of fire tomorrow.

Yeah i was just reading that. Probably has to do with back to school sales and more school's are forcing kids to have laptops and such and they are making the parents buy them. Also PC gaming is once again big and beating console software sales.
Edit i actually think this is scaring Microsoft and Sony
http://www.dsogaming.com/news/report-microsoft-sony-pressuring-ubisoft-for-30fps-on-pc/