[Xbitlabs]AMD Excavator Core May Dramatic Performance Increases.

[Exophase]

IIRC Steamroller is to be launched in Q114, and we are yet to hear anything of that processor. And instead of bringing attention to the soon-to-be launched product, AMD is trying to bring attention to... Excavator, a product which is at least 6 quarters away.

Gentlemen, get ready for another deception once Steamroller hit the review sites.

AMD isn't trying to bring attention to Excavator. They made a commit to GCC and Anton Shilov is running with it like wild. The whole article is really dumb, both in that he thinks AVX2 support means 256-bit execution units (when AVX support didn't) and that he thinks that this alone would mean a huge performance boost for the CPU.

 

[Ventanni]

I'm surprised you didn't think of the IPC jump Conroe had over Prescott.

That was my first thought, too.

 

[eyeofcore]

Don't know if I should comment on that or just report this and move along

You could thought it would only back fire at you and bite your rear end...

 

[Kenmitch]

If AVX2 really takes off, it will result in a large number of toasty Haswells from what I'm hearing...

Haswell + stock settings + stock cooler + crappy case = Throttle with Linpack 11.

 

[krumme]

Intel is cruising now and could release 8 core haswell for consumer enthusiast if they had to.
I dont care if excavator is a bomb giving 8 core haswell speed.
At that time nearly 2 years from now its nearly skylake time and Intel could release a 16 core monster if they had to. Hell they could even make it a real desktop processor. Lol.

(Edit: at that time we are probably at jaguar derivative 2 at 16nm. Surely at 8 cores with 20% plus ipc and 50% freq. One have to wonder if more is needed for the most of the market.)

 

[tential]

I'm just in the "I'll believe it when I see it camp."

I always hear how AMD is going to change the game. Then it doesn't happen.

Not to mention we're talking too different power usages when comparing intel/AMD in desktops.

Not saying I don't believe they could do it because well, Conroe blew our socks off. But just because of some leaked stuff that has nothign at all concrete and 0 information, I'm not going to jump no some AMD stock now thinking "Ha, intel doesn't know what's coming!"

 

[MisterMac]

I'm worried there hasn't been more leaks regarding SR.

The overall APU Should be QUITE tasty for the low end "enthusiast".
They should be feeding small teasers about performance and creating buzz.


...why is there nothing?

 

[24601]

I'm worried there hasn't been more leaks regarding SR.

The overall APU Should be QUITE tasty for the low end "enthusiast".
They should be feeding small teasers about performance and creating buzz.


...why is there nothing?

Because it's super late. As always.

 

[Centauri]

If I take anything from these articles, it's that AMD is confident enough to even mention 14nm at this point. That's an excellent sign - I'd begun to wonder if even 20nm was still a reasonable expectation in the midterm.

 

[rtsurfer]

I really hope AMD releases this on AM3+. But I guess it's more likely they will release on a new FMx socket

I really hope not. I wish that they would move to a new socket.

 

[Shivansps]

I'm surprised you didn't think of the IPC jump Conroe had over Prescott.

That has to be the bigger jump ever.

 

[tential]

That has to be the bigger jump ever.

It was mind shattering review sites went insane. Then the OC potential too just was mind blowing given how well it was performing. Athlon 64 was just taking over then intel was like "Nope....". Since then, intel learned what happens when you slack off.

I still have my Athlon 64 machine and it still runs.

If AMD could Conroe intel that'd be interesting but I think we're just at a so much more advanced state. More power while great isn't what people are solely looking for. It has to be at the correct power consumption.

AMDs biggest problem is trying to compete with Nvidia, and Intel at the same time while also even competing with the QualComm/Apple (as they're trying to get into tablets too).

I wouldn't wish that on anyone.

 

[OBLAMA2009]

Try the 3rd option.

ma taughts ezackly

 

[sniffin]

I'm worried there hasn't been more leaks regarding SR.

The overall APU Should be QUITE tasty for the low end "enthusiast".
They should be feeding small teasers about performance and creating buzz.


...why is there nothing?

AMD not saying a word has historically been a better thing.

 

[Ancalagon44]

If I take anything from these articles, it's that AMD is confident enough to even mention 14nm at this point. That's an excellent sign - I'd begun to wonder if even 20nm was still a reasonable expectation in the midterm.

AMD was also confident bulldozer IPC would be increased, just saying.

I must say, its been a while since AMD screwed up - Richland performed as expected, and Jaguar is pretty good - but when it comes to promises of major performance increases, I'm taking a "I'll believe it when I see it approach."

 

[Exophase]

AMD was also confident bulldozer IPC would be increased, just saying.

I'm a little skeptical about this. To the best of my knowledge it was one person, JF-AMD, claiming it (and not in an especially official capacity like via presentation material). I know he used the excuse that the engineers promised him this and failed to deliver but I wonder if he didn't misunderstand something. His blog and forum posts were rife with basic misunderstandings of technical data, including how his own company's processors worked.

 

[mikeymikec]

Where are you drawing that conclusion from? None of those graphs have data for any of the K6 series.

My mistake. I read "3DNow!" and my brain inserted "K6 with 3DNow"

- edit - I've finally found a benchmark that includes an early Athlon and a K6-2:
http://www.tomshardware.com/reviews/mother-cpu-charts-part-2,944-19.html

Pretty big difference there! The K6-2s are mostly missing from the benchmark results preceding and following that one. A quick calculation (which may be incorrect) suggests that clock-for-clock the Athlon was twice as fast as the later K6-2s, but a broader range of benchmarks would help clarify the result. My calculation for the WinRAR results draw a different picture though (my calculation was time in minutes * K6 clock speed / athlon clock speed), they're almost the same in that respect.

I'm surprised you didn't think of the IPC jump Conroe had over Prescott.

I thought of it later:
http://forums.anandtech.com/showpost.php?p=35625432&postcount=36

Perhaps the jump to SB from Conroe stands out in my mind more because a) it was more recent and b) it was basically a coup de grace against AMD and dispelled any notion that AMD was going to catch up at any point soon, because it showed Core i3 SBs spanking the best of AMD's line in a lot of cases.

Because it's super late. As always.

Probably fab problems again.

As much as I liked AMD (and in the days of AMD CPUs in nForce boards, I loved them! Cheap, decent performers with reasonable chipset QA!), and I appreciate the benefits of competition, Intel would need to screw up big time (ie. miss an enormous opportunity) for AMD to regain the lead on desktop CPU performance (probably mobile too, I don't pay much attention to it). The IGP gaming market is a bit of a bizarre one to aim for IMO. Perhaps that and the ATI markets are the best that AMD can hope for right now?

 

[anthd56]

Precisely.

7.7% faster at equal clocks in applications.
3.8GHz-4GHz FX4300 is a bit slower than a 4.2-4.3GHz FX4170.
3.2GHz quad-core Phenom II matches a 3.8-4.0GHz 2M/4C Piledriver. Time to catch up to Phenom II level IPC Steamroller.

Man stop comparing a quad core to a 2M/4T, the FX4300 is not a real quad core. The 955 is more comparable to the FX8320 which has 4 modules and around the same end of the market on release. The FX4300 is more comparable to the PII X2 and FX6XXX to the PII X3 imo. A 2M/4T CPU performing close to a PII Quad is a good thing.

But I agree, AMD does need more IPC, but unless they can do a Prescott to Conroe like jump in IPC, I doubt they'll ever catch up to Intel. Another thing is power consumption, Its ridiculous how the FX8300s consume 125W and on the Intel side they have quads under 84W that are way faster.

 

[nemesismk2]

i am excited about this AMD Excavator news but i want the power consumption to drop a bit because AMD can't always increase power consumption just to get better performance.

 

[mikeymikec]

i am excited about this AMD Excavator news but i want the power consumption to drop a bit because AMD can't always increase power consumption just to get better performance.

They might stop when the cooling system uses/costs more than the rest of the machine

 

[Vesku]

Power efficiency is going to be quite a bit in TSMC and GlobalFoundries ball court, yes? Of course AMD can greatly help by de-Pentium4ing their design and targeting more reasonable clocks. Remember that die shot that looked quite nice in terms of improving upon Bulldozer/Piledriver? Even more likely that was a peek at Excavator, it seemed to have dual or quad 256 bit FPUs.

This news is interesting in that it means AMD probably isn't completely retreating into the cat family and 64 bit ARM.

Edit- I believe I was thinking of this thread, http://forums.anandtech.com/showthread.php?t=2322010&highlight=steamroller , when mentioning the potential excavator die shot.

 

[nenforcer]

This news is interesting in that it means AMD probably isn't completely retreating into the cat family and 64 bit ARM.

There is still too much money in the x86 market for them to retreat. There will always have to be an alternative to Intel both in the consumer and server market. Some people just like the underdog or just "good enough" CPU performance of AMD.

It's just that now there is a third party in low power ARM chips in the mobile and soon to be server space. AMD is arguably better positioned by diversifying their product portfolio while Intel just continues to push X86 everywhere.

 

[BallaTheFeared]

Man stop comparing a quad core to a 2M/4T, the FX4300 is not a real quad core.

lol?

It's more a quad core than a 2c/4t chip.

Saying it's 2c/4t makes it seem like Intel, high per core, but low core count but it isn't. It has the same poor performance in dual threaded applications as it does in 4+ threaded ones.

 

[Piroko]

I'm surprised you didn't think of the IPC jump Conroe had over Prescott.

Technically, Penryn was Conroes predecessor. The IPC increase of Conroe over Prescott was dramatic, but Conroe also dropped the clocks quite low.

AMD not saying a word has historically been a better thing.

I laughed harder than I should have.

My mistake. I read "3DNow!" and my brain inserted "K6 with 3DNow"

- edit - I've finally found a benchmark that includes an early Athlon and a K6-2:
http://www.tomshardware.com/reviews/mother-cpu-charts-part-2,944-19.html

Yeah, the K7 performance increase back then was quite dramatic over K6-2 and even K6-3, especially in games (or FPU heavy programs).

 

[Xpage]

I am more worried that there are no SR leaks on information, i thought it was a Q4 release (ie benches) and Q1 rollout.

I think SR will have a better IPC incerase than XC as it will be incorporating things than SB has such as the uops cache that i think does wonders. Hopefully they have the L3 cache latency issue finally fixed in this too.