AMD CPU speculation... and expert conjecture

[szatkus]

Phenom was better quad-core.
Core 2 Quad was better core.
Phenom II was great because it was small, cheap and clocks were high. First i7 were released little earlier, but it was expensive. From perspective of average consumer Phenom II was really good, but not the best.

 

[-Fran-]

Pardon my stupidity but isnt x^3 a parabolic and not a straight line? Its been soo long since i had jr high algebra that i simply dont remember.

Can anyone help me.

Phenom was better quad-core.
Core 2 Quad was better core.
Phenom II was great because it was small, cheap and clocks were high. First i7 were released little earlier, but it was expensive. From perspective of average consumer Phenom II was really good, but not the best.

Basically, Phenom II, step C was the best. They were the ones who could get near 4Ghz with little effort making them the sweetest deal when the QX CPUs were stupid high in price.

BUT!

Talking about the past won't fix the present.

In my awesome and incredible opinion, "WTFTECH" can speculate with the earnings call announcement all they want as long as it makes sense. They don't need to be 100% right, just make a sound argument to shed some light into a very dark tunnel. I take that as a "half full glass", with all the salt we need to take

Cheers!

 

[esrever]

I remember toms did an analysis of c2q and phenom and more and it did show that phenom scaled to 4 thread better than c2q. Phenom just had too low clocks.

 

[szatkus]

In my awesome and incredible opinion, "WTFTECH" can speculate with the earnings call announcement all they want as long as it makes sense. They don't need to be 100% right, just make a sound argument to shed some light into a very dark tunnel. I take that as a "half full glass", with all the salt we need to take

Cheers!

At WCCF I only look for sources. Because they are really good at finding leaks. But their "technical" analysis are usually plain stupid. Even Fudzilla is better, and thet don't even pretend to be competent.

 

[szatkus]

I remember toms did an analysis of c2q and phenom and more and it did show that phenom scaled to 4 thread better than c2q. Phenom just had too low clocks.

At the time AMD's 65nm was really bad.

 

[juanrga]

It took them 6 long years, but Finally they listened.

http://wccftech.com/amd-earnings-call/

What Lisa Su said was "we will increase our R&D investments in enterprise, embedded, and semi-custom." The rest was invented/guesses by WCCFTECH.

Moreover during the rest of the talk she mentioned that AMD will reduce its "concentration in consumer PCs".

Think you are glossing over a key part of that R&D budget. While, in this case, wccftech is making some inferences from the announcement...they are not unjustified in doing so.

I am rather sure the increase in R&D is coming from the clients are paying AMD to develop the semicustom designs based in K12 and Zen. And those customers prefer smaller, efficient, and cheap cores (Jaguar style) rather than big, powerful, and expensive cores.

 

[jdwii]

I think its a great thing myself^, the fact Amd even has fans that really want them to make things is incredible.

 

[con635]

Had to a catch with the gtx970
http://www.extremetech.com/extreme/198214-198214

 

[juanrga]

^^^^ let put a spotlight on the above post, amd is using other people's money to develop next gen server parts and that is a bad thing????

The post was simply descriptive (what is happening) not interpretative (is good or bad?).

The content of the post was emphasizing that WCCFTECH was trying to sell us another Baeca.

 

[jdwii]

Had to a catch with the gtx970
http://www.extremetech.com/extreme/198214-198214

Been reading on it could either be a design flaw or some other driver/firmware mistake since some people aren't having the issues.

What i find even more funny is how people are acting like the card sucks now when its still even with a 290X according to techpowerup. When Nvidia fixes the issue it will over perform better, might explain why the card was losing by a few % in 4K compared to a 290X however.

Darn it Nvidia, hopefully by the time i get a new video card this problem is fixed. Still waiting for the best 200 watt card for 350$ or so. I'll go Amd i'm currently on their drivers again right now with my llano laptop and it just gets me mad they don't have adaptive v-sync or a smaa/fxaa injector. I use that in nearly all my games, except AOE3 which for some reason doesn't work with adaptive v-sync with Nvidia.

 

[con635]

^^The problem isn't general performance from what Ive read, its a horrible stutter in some situations where more than 3.25gb of vram is used, sort of makes it a benchmark queen although I understand you, I would still buy one if price was right, whats interesting is what nvidia will do if its hardware a recall or rebadge it as 3gb and not use 1gb maybe via drivers? The internet pitchforks are watching closely.

 

[con635]

http://www.pcper.com/news/Graphics-Cards/NVIDIA-Responds-GTX-970-35GB-Memory-Issue

 

[-Fran-]

I wonder what nVidia is going to say this time around, hahaha.

https://www.youtube.com/watch?v=ZQE6p5r1tYE

That looks BAD.

Cheers!

 

[esrever]

I wonder what nVidia is going to say this time around, hahaha.

https://www.youtube.com/watch?v=ZQE6p5r1tYE

That looks BAD.

Cheers!

That looks bad but I don't think its the same bug, looks like completely messed drivers with the flickering.

 

[blackkstar]

Woah, I thought this thread was dead. There's some sort of bug in the forum for me where if you go to page 348, the next pages don't show up. I almost stopped coming here.

Some of you seem surprised by AMD cutting things from their road maps. They are directly in the middle of a transitioning period between Bulldozer and Zen/K12. They are going to cut things. Fottemberg at S|A forums, who has a good history of being right, is calling for 14nm CPUs from AMD to compete with 14nm CPUs from Intel in Q5 2015 or Q1 2016.

Carrizo is looking at a Q2 2015 release. I don't know if that means we're getting a high end x86 CPU at that time, but it seems like ignoring the fact that AMD could just be abandoning those plans so they don't end up with a ton of Carrizo inventory left over once Zen/K12 comes out is a huge mistake.

And we should all know by now that of coures AMD R&D is going to go down. The entire purpose of semi-custom is to reduce spending on R&D while having it minimally affect your products negatively.

Carrizo for HEDT in FM2+ would need a new die, like you all said. After semi-custom, those sort of things are going to be a lot easier. I'm expecting AMD to release a solid x86 product that works well in HEDT. Except it won't be designed for HEDT, it'll be table scraps from Enterprise, Server, HPC, etc. Also, look at the timetable bewteen Excavator being released and Zen coming out. Releasing Excavator chips in a lot of markets, only to be made obsolete a half a year later seems a little counter-intuitive. AMD just went all in for mobile with Carrizo because they wanted to recover as much as possible while spending as little as possible.

 

[szatkus]

Q5 2015

 

[Reepca]

Q5 2015

First thing I noticed too. I assume the "Q" in Q1, Q2 etc means "quarter"...

Speaking of which, do business quarters reflect the actual Calendar? Where Q1 would be Jan through end of March, Q2 would be April through end of June, etc?

So someone saying Carrizo will be released "sometime in Q2" could be any time from April to end of June?

 

[jdwii]

http://kotaku.com/then-and-now-five-generations-of-radeon-graphics-compa-1680378547

Pretty nice article for you guys

Then And Now: Five Generations Of Radeon Graphics Compared

 

[juanrga]

Some of you seem surprised by AMD cutting things from their road maps. They are directly in the middle of a transitioning period between Bulldozer and Zen/K12. They are going to cut things. Fottemberg at S|A forums, who has a good history of being right, is calling for 14nm CPUs from AMD to compete with 14nm CPUs from Intel in Q5 2015 or Q1 2016.

The main reason why AMD is canceling products is the lack of funds and their perennial myopie when planing long-term strategy. Being stuck to the 28nm node means that engineers have to look elsewhere to reduce power consumption if want to be barely competitive against Broadwell. Then they rely on stuff such as HDL, and cutting L2 caches to reduce area and power consumption. HDL also reduces frequencies.

As consequence, they cannot use Excavator for APUs above 35W and the previous 45W/65W Carrizo APUs get canceled because would be slower than Kaveri parts. The same reason why Excavator is not coming to FX (AM3+).

There is something interesting about Carrizo that all of us are not paying enough attention. The socket compatibility between Carrizo and Carrizo-L, aka the convergence between Excavator (which is a smaller core than Steamroller) and Puma+ (which is an improved Jaguar).

This seems to confirm that AMD will fuse big and small cores into a single line (K12/Zen). Thus, the new cores will be sized somewhat between both extremes. This means that AMD abandons any pretension to compete with Intel/IBM/ARM-players on big cores. This explains why K12 is targeting dense servers instead high-performance servers, and why Cray didn't select AMD for its future supercomputers.

One has to remember that semicustom/embbeded clients are paying AMD to develop K12/Zen, but those clients don't want big, powerful, and expensive cores; but small, efficient, and cheap cores (Jaguar style).

When Fottemberg said that K12/Zen was being made on TSMC 16nm, I said that was being made on Glofo/Samsung 14nm node (something has been confirmed lately). I will say here that he is being optimist about schedule and that K12 will be not ready before Q1 2016, with Zen coming latter.

Fottemberg contacts said him that AMD prepares 20-core FX CPUs. This confirms that AMD is again trying to hide its deficiency on single thread performance with "moar cores". Most HEDT users will prefer 4/6/8 stronger cores (Skylake) rather than 20 weaker Zen cores.

All points to AMD repeating Bulldozer mistake once again!!!

 

[esrever]

http://kotaku.com/then-and-now-five-generations-of-radeon-graphics-compa-1680378547

Pretty nice article for you guys

Then And Now: Five Generations Of Radeon Graphics Compared

290x is now 2% faster than the 980ti according to that.

 

[cdrkf]

Actually it's saying 290x is 2% faster than the 780ti.... Apparently it's still 15% slower than a 980 (though I'll bet that's at 1080p... I'm still convinced the 970 and 980 cards aren't going to age well with their puny 256bit memory interface, which becomes more obvious as resolution goes up).

 

[cemerian]

Some of you seem surprised by AMD cutting things from their road maps. They are directly in the middle of a transitioning period between Bulldozer and Zen/K12. They are going to cut things. Fottemberg at S|A forums, who has a good history of being right, is calling for 14nm CPUs from AMD to compete with 14nm CPUs from Intel in Q5 2015 or Q1 2016.

The main reason why AMD is canceling products is the lack of funds and their perennial myopie when planing long-term strategy. Being stuck to the 28nm node means that engineers have to look elsewhere to reduce power consumption if want to be barely competitive against Broadwell. Then they rely on stuff such as HDL, and cutting L2 caches to reduce area and power consumption. HDL also reduces frequencies.

As consequence, they cannot use Excavator for APUs above 35W and the previous 45W/65W Carrizo APUs get canceled because would be slower than Kaveri parts. The same reason why Excavator is not coming to FX (AM3+).

There is something interesting about Carrizo that all of us are not paying enough attention. The socket compatibility between Carrizo and Carrizo-L, aka the convergence between Excavator (which is a smaller core than Steamroller) and Puma+ (which is an improved Jaguar).

This seems to confirm that AMD will fuse big and small cores into a single line (K12/Zen). Thus, the new cores will be sized somewhat between both extremes. This means that AMD abandons any pretension to compete with Intel/IBM/ARM-players on big cores. This explains why K12 is targeting dense servers instead high-performance servers, and why Cray didn't select AMD for its future supercomputers.

One has to remember that semicustom/embbeded clients are paying AMD to develop K12/Zen, but those clients don't want big, powerful, and expensive cores; but small, efficient, and cheap cores (Jaguar style).

When Fottemberg said that K12/Zen was being made on TSMC 16nm, I said that was being made on Glofo/Samsung 14nm node (something has been confirmed lately). I will say here that he is being optimist about schedule and that K12 will be not ready before Q1 2016, with Zen coming latter.

Fottemberg contacts said him that AMD prepares 20-core FX CPUs. This confirms that AMD is again trying to hide its deficiency on single thread performance with "moar cores". Most HEDT users will prefer 4/6/8 stronger cores (Skylake) rather than 20 weaker Zen cores.

All points to AMD repeating Bulldozer mistake once again!!!

I have been thinking about the 20 small core fx rumor,So if it's a new architecture with 20 core the single thread performance will be insanely small, but you have all forgotten amd co-developed that threading tech that allows to use many small cores to make 1 big virtual core(the oposite of smt kind of) can't remember what was it called, was quite some time ago but still if that works as good as they promised there may still be a bit of hope(realistically none)

 

[logainofhades]

I think its a great thing myself^, the fact Amd even has fans that really want them to make things is incredible.

I am an AMD fan, but I also see the reality that AMD, right now, is not worth getting on the cpu side, unless budget is extremely low, and you live near a Microcenter. I blame poor management, for the problems they are in. Current management seems to be trying to dig AMD out of the hole it is in. I hope they do it. I want real competition again.

 

[juanrga]

Some of you seem surprised by AMD cutting things from their road maps. They are directly in the middle of a transitioning period between Bulldozer and Zen/K12. They are going to cut things. Fottemberg at S|A forums, who has a good history of being right, is calling for 14nm CPUs from AMD to compete with 14nm CPUs from Intel in Q5 2015 or Q1 2016.

The main reason why AMD is canceling products is the lack of funds and their perennial myopie when planing long-term strategy. Being stuck to the 28nm node means that engineers have to look elsewhere to reduce power consumption if want to be barely competitive against Broadwell. Then they rely on stuff such as HDL, and cutting L2 caches to reduce area and power consumption. HDL also reduces frequencies.

As consequence, they cannot use Excavator for APUs above 35W and the previous 45W/65W Carrizo APUs get canceled because would be slower than Kaveri parts. The same reason why Excavator is not coming to FX (AM3+).

There is something interesting about Carrizo that all of us are not paying enough attention. The socket compatibility between Carrizo and Carrizo-L, aka the convergence between Excavator (which is a smaller core than Steamroller) and Puma+ (which is an improved Jaguar).

This seems to confirm that AMD will fuse big and small cores into a single line (K12/Zen). Thus, the new cores will be sized somewhat between both extremes. This means that AMD abandons any pretension to compete with Intel/IBM/ARM-players on big cores. This explains why K12 is targeting dense servers instead high-performance servers, and why Cray didn't select AMD for its future supercomputers.

One has to remember that semicustom/embbeded clients are paying AMD to develop K12/Zen, but those clients don't want big, powerful, and expensive cores; but small, efficient, and cheap cores (Jaguar style).

When Fottemberg said that K12/Zen was being made on TSMC 16nm, I said that was being made on Glofo/Samsung 14nm node (something has been confirmed lately). I will say here that he is being optimist about schedule and that K12 will be not ready before Q1 2016, with Zen coming latter.

Fottemberg contacts said him that AMD prepares 20-core FX CPUs. This confirms that AMD is again trying to hide its deficiency on single thread performance with "moar cores". Most HEDT users will prefer 4/6/8 stronger cores (Skylake) rather than 20 weaker Zen cores.

All points to AMD repeating Bulldozer mistake once again!!!

I have been thinking about the 20 small core fx rumor,So if it's a new architecture with 20 core the single thread performance will be insanely small, but you have all forgotten amd co-developed that threading tech that allows to use many small cores to make 1 big virtual core(the oposite of smt kind of) can't remember what was it called, was quite some time ago but still if that works as good as they promised there may still be a bit of hope(realistically none)

I think you must be referring to this FUD

http://www.theinquirer.net/inquirer/news/1009078/reverse-hyperthreading-exist

There are different academic/industrial attempts to run a single thread on two or more cores and the contributions of AMD to this are easily summarized: zero.

I wrote a recent post about single-thread performance on RWT. I will copy and paste part of it:

VISC (doesn't stand for Virtual Instruction Set Computing). A clustered architecture (up to scheduling) plus internal new ISA optimized for threadlets. Single thread performance follow expected law IPC ~ sqrt(N) with N number of execution cores.

This approach retains usual front-end bottlenecks, doesn't improve branching, and doesn't improve the instruction window, which will limit the wide.

DCE (Dual Core Execution). This approach combines two IOE cores to form a big OOOE core. Head core executes a skeleton of the single thread to provide prefetching and accurate branch prediction for the tail core that runs the full thread. Performance gains are as expected. Up to 200% increase with average of about 41%.

Fusion/Federation. Similar to the last. Two or more simple cores are combined to do a bigger core. However, here the execution model is not runahead but the fused cores work as a monolithic core by combining caches, decoders,... Performance scales similar to monolithic core of the same wide.

TLS (Thread Level Speculation). Breaks a single thread into threadlets that are then run on different cores. Similar to VISC if the split is HW based, but can improve branch prediction via redundant execution. Performance follows the usual square law. On a 64-core [strike]pink unicorn[/strike] manycore, a single thread can be speed up to 8 times.

The last is also known as SpMT (Speculative MultiThreading). In any case, the technologies are not developed enough for commercial usage and the chances that AMD has implemented something similar on Zen are epsilon when epsilon tends to zero.

The most probable implementation for Zen is a traditional CMP superscalar OOOE muarch. I even doubt AMD will implement SMT.

I have been discussing the 20-core FX rumor with Fottemberg and shared with him some math. A 20-core die is compatible with small Zen cores (Jaguar style) having about 40% higher IPC than Jaguar/Piledriver.

 

[8350rocks]

Some of you seem surprised by AMD cutting things from their road maps. They are directly in the middle of a transitioning period between Bulldozer and Zen/K12. They are going to cut things. Fottemberg at S|A forums, who has a good history of being right, is calling for 14nm CPUs from AMD to compete with 14nm CPUs from Intel in Q5 2015 or Q1 2016.

The main reason why AMD is canceling products is the lack of funds and their perennial myopie when planing long-term strategy. Being stuck to the 28nm node means that engineers have to look elsewhere to reduce power consumption if want to be barely competitive against Broadwell. Then they rely on stuff such as HDL, and cutting L2 caches to reduce area and power consumption. HDL also reduces frequencies.

As consequence, they cannot use Excavator for APUs above 35W and the previous 45W/65W Carrizo APUs get canceled because would be slower than Kaveri parts. The same reason why Excavator is not coming to FX (AM3+).

There is something interesting about Carrizo that all of us are not paying enough attention. The socket compatibility between Carrizo and Carrizo-L, aka the convergence between Excavator (which is a smaller core than Steamroller) and Puma+ (which is an improved Jaguar).

This seems to confirm that AMD will fuse big and small cores into a single line (K12/Zen). Thus, the new cores will be sized somewhat between both extremes. This means that AMD abandons any pretension to compete with Intel/IBM/ARM-players on big cores. This explains why K12 is targeting dense servers instead high-performance servers, and why Cray didn't select AMD for its future supercomputers.

One has to remember that semicustom/embbeded clients are paying AMD to develop K12/Zen, but those clients don't want big, powerful, and expensive cores; but small, efficient, and cheap cores (Jaguar style).

When Fottemberg said that K12/Zen was being made on TSMC 16nm, I said that was being made on Glofo/Samsung 14nm node (something has been confirmed lately). I will say here that he is being optimist about schedule and that K12 will be not ready before Q1 2016, with Zen coming latter.

Fottemberg contacts said him that AMD prepares 20-core FX CPUs. This confirms that AMD is again trying to hide its deficiency on single thread performance with "moar cores". Most HEDT users will prefer 4/6/8 stronger cores (Skylake) rather than 20 weaker Zen cores.

All points to AMD repeating Bulldozer mistake once again!!!

This does not reflect any of what I am hearing. Honestly...where do you get this stuff? AMD has said nothing of a 20 core CPU...maybe a 20 CU APU. 6-8 serial cores and 12-14 GCN modules....

That would be much closer to what I have heard about APUs down the line....still does not jive with CPU rumors though...