Some benches on the upcoming 45nm AMD

[deanjo]

http://img.inpai.com.cn/2008/8/5/0f50e1a7-5451-466f-8b09-a4c2fae025f0.gif

http://img.inpai.com.cn/2008/8/5/74c966f3-d032-4f81-b3fc-82c11f710554.gif

http://img.inpai.com.cn/2008/8/5/7befb69f-ce9a-4b92-8340-72068fc5e960.gif

http://img.inpai.com.cn/2008/8/5/57e23f7f-3ad2-4024-baba-52bc32f8dbaf.gif


Quake Wars

http://img.inpai.com.cn/2008/8/6/959bae4f-bf3a-40c7-91fc-13e6d7255ead.gif

Crysis

http://img.inpai.com.cn/2008/8/6/9061c6a0-7a6d-439a-8054-bb706d8d543a.gif

BluRay (Pirates O.T.C.)

http://img.inpai.com.cn/2008/8/6/b6f5151d-b9e2-4fdd-a37d-446b9ad8d817.gif

Idle Power

http://img.inpai.com.cn/2008/8/6/24dfb59e-41b0-4b39-a547-bac80f883a83.gif

Full Load Power

http://img.inpai.com.cn/2008/8/6/2b75a295-75d3-4273-ac8b-0a287f1301dc.gif

[deanjo]

http://xtreview.com/images/Phenom%2045%20nm%20benchies%2002.jpg
http://xtreview.com/images/Phenom%2045%20nm%20benchies%2001.jpg
http://xtreview.com/images/Phenom%2045%20nm%20benchies%2003.jpg
http://xtreview.com/images/Phenom%2045%20nm%20benchies%2004.jpg

I'm looking forward to see how PTS results would be.

[d2kx]

I can't speak for the second comment, but the first comment includes (very) old results with an early stepping, so be careful with your conclusions. But the reduced power consumption is pretty amazing to say at least.

[deanjo]

I wouldn't say two months is all that old.

[Ex-Cyber]

What's being measured in the POV-Ray benchmark? Pixels per second?

[deanjo]

4 Ghz from a 2.2 Ghz proc on air :O

http://www.coolaler.com.tw/toppc/AMD45NM/4G.JPG

[duby229]

4 Ghz from a 2.2 Ghz proc on air :O

That's impressive, but he's running it at 1.6v on a 45nm process. I wouldnt run a 90nm chip at that voltage let alone a 65nm chip or even worse 45nm. He probably disabled every service that he could, then booted it up, and started cpuz then took a snapshot as fast as he could before it locked up on him.

Still though If it runs at 3.4ghz or 3.5ghz on air at reasonable volts and temps that would be mighty impressive. From the looks at that snapshot it may well be possible to do just that.

[MetalheadGautham]

Any benchmarks on the compilation times in Gentoo ?

[mattmatteh]

i would be nice to include how many watts that using. i know some users dont mind power wasting cpus but i am interested in power saving cpus. or at the least proceesing power / watt. and idle power usage.

[deanjo]

i would be nice to include how many watts that using. i know some users dont mind power wasting cpus but i am interested in power saving cpus. or at the least proceesing power / watt. and idle power usage.

While I can appreciate the "green" concern, truth is that you will probably save more per year changing 2 incandescent bulbs over to led or florescent in your house then picking a 65 watt over a 125 watt processor. Performance per watt is also very determinant on the software you are running and their optimization for a particular processor. There is also other factors to consider as well such as logic chipsets as well. While the savings can be huge in a server farm / data center scenario where systems operated under heavy load 24/7, chances are in 2-3 years you will be swapping out the system anyways and the savings to be had by a lower power consumption processor will be lost by the extra initial purchase price of the lower consumption processor.

[mattmatteh]

its not so much as being green as the heat that is generated. its nice during the winter, but not the summer. i dont have any air conditioning here, and dont think that should be reqired to compute.

matt

[deanjo]

its not so much as being green as the heat that is generated. its nice during the winter, but not the summer. i dont have any air conditioning here, and dont think that should be reqired to compute.

matt

OK I can appreciate that but keep in mind that there are probably more BTU's coming from the drives. video cards. chipset and power supply then just the cpu.

[mattmatteh]

agreed. i look at all the components, not just the cpu. i find the reviews that show the power usage from the wall plug the best.

matt

[deanjo]

Another write up from the techreport today.


AMD overclocks 45nm Phenoms to 4GHz and beyond
by Cyril Kowaliski (ckowaliski@techreport.com) — 1:47 PM on November 20, 2008
According to AMD's latest roadmaps, 45nm Phenom II processors are just a few weeks away from launch—they'll materialize at the Consumer Electronic Show in early January. Not a bad time to start gauging the overclocking potential of these chips, right?
AMD took care of that itself at an event in Austin this morning, where it showcased four overclocked Phenom II systems. TR editor-in-chief Scott Wasson was on the scene, and while AMD didn't allow attendees to take pictures, he jotted down a few details.
The slowest system—cooled with a heatsink and fan—managed to reach just under 4GHz with a 1.55V core voltage. With liquid cooling, AMD successfully pushed a 45nm Phenom II in another machine just over the 4GHz mark. That required kicking up the CPU voltage to 1.6V, however. For the other two systems, AMD took out the big guns. One was strapped to a phase-change cooler and reached the mid-4GHz range at 1.7V, while the fastest system managed to break the 6GHz barrier using liquid nitrogen (which kept the core temperature down to a chilly -185°C).
Considering the latest leaked roadmaps suggest the fastest Phenom II launch CPU will run at 3GHz, those are pretty decent overclocks. We should note that AMD was using Crysis to test stability, however, and that only stresses a couple of cores at best.


http://techreport.com/discussions.x/15927

[psycho_driver]

Well, the real benchmarks are hitting the web now, and congratulations are in order to AMD. While still a generation behind Intel, they do now have a serious competitive product for the entry level quad core market.

Pricing trends over the next couple of months will now determine whether my next system build is on an AMD or Intel platform, whereas previously I assumed I would be choosing within the Intel product lineup exclusively.

[deanjo]

Well, the real benchmarks are hitting the web now, and congratulations are in order to AMD. While still a generation behind Intel, they do now have a serious competitive product for the entry level quad core market.

Pricing trends over the next couple of months will now determine whether my next system build is on an AMD or Intel platform, whereas previously I assumed I would be choosing within the Intel product lineup exclusively.

Ya they look promising. Personally I think I'll be holding off until the AM3 945's are out. Buying a processor such as the 940 that is limited to AM2+ boards doesn't leave a lot of flexibility in future MB upgrade options. Until then my 9850's will do just fine. It will also give a bit extra time for the refinement of the manufacturing process.

[MetalheadGautham]

Honestly, I think they suck. NO AM3, no HT 3.0 and NO DDR3.

That I could forgive, but for the fact that clock-to-clock, they lag behind even the Q9x50 processors. Phenom II 940 just equals Core2Quad Q9450.

And Core i7 is waaaay ahead of it. Multimedia performance is so low compared to Yorkfield and Bloomfield. And AMD has nothing to combat Intel's HT Technology.

But I definitely think the Phenom II CPUs are a stepping stone for AMD. Reminds me of the time RV600 came out. Were no match for nVidia's higher end, but the lower end and midrange performed just fine. Then RV700 came and changed the whole situation. If things go well, AMD Bulldozer may overthrow Intel SandyBridge, but thats only IF AMD manages toreach 32nm before Intel does.

[Kano]

Well I would bench it if I would get a free cpu, board + gfx card from AMD ;)

[psycho_driver]

Multimedia performance is so low compared to Yorkfield and Bloomfield.

That discrepancy is overblown by the synthetic benchmarks. You don't see that big of a difference in the real world benchmarks.

The introduction of the Phenom IIs really made my new upgrade a hard decision. I wasn't interested at this point in i7 stuff, because the motherboards and memory are still at too high of a premium. Therefore I had to basically choose between a p45 platform with not much of a future in regards to upgradability, or a 790gx platform that would underperform now but potentially be able to accept a higher clocking, more affordable phenom II at some point down the road that may end up outperforming any quad lga775 offerings Intel releases from here on out.

After days of product comparisons I ordered my p45 componants yesterday. Probably the biggest deciding factor is that there were a lot of established, quality motherboards with the p45 chipset around the price I was looking to pay compared to what's available with 790gx/750sb in the same range.

Maybe AMD will really be back in the race a couple of years from now when I need to build my next system.

[Kano]

I would prefer a Q9300 or Q9550 + P45 over anything from AMD if you want to buy something new. I could not test the new AMD system, but I know what works now...

[highlandsun]

Honestly, I think they suck. NO AM3, no HT 3.0 and NO DDR3.


Eh? HT3.0 is already part of AM2+.

And the AM2+ processors all have DDR3-capable memory controllers, it's just that nobody has made any DDR3 motherboards for them. And as far as I can tell, there's really not much point at the moment.


And Core i7 is waaaay ahead of it. Multimedia performance is so low compared to Yorkfield and Bloomfield. And AMD has nothing to combat Intel's HT Technology.


i7 is definitely far ahead. Nothing else to say about that.

As for multimedia, I'm still suspicious of the compiler technology; architecturally K10's SSE support looks really good on paper. It seems to me that we just haven't seen code that was generated properly to take full advantage of it. And HT really doesn't buy anything if you can keep your active cores sufficiently fed that they have fewer stalls.

[deanjo]

Eh? HT3.0 is already part of AM2+.

And the AM2+ processors all have DDR3-capable memory controllers, it's just that nobody has made any DDR3 motherboards for them. And as far as I can tell, there's really not much point at the moment.

Actually the AM2+ Phenom II's have the DDR3 mem controllers disabled. You will never see a AM2+ Phenom II with DDR3. I actually would think that the AM2+ Phenom II's are actually early revisions with faulty DDR3 controllers but their DDR 2 controllers worked fine. The AM2+ Phenom II's will never work in a AM3 board because of this (same with any other AM2+ processor). Your right though HT 3 has been around for a long time though already. Your also correct that DDR3 at this point doesn't offer much, if any, in real world performance.



i7 is definitely far ahead. Nothing else to say about that.

As for multimedia, I'm still suspicious of the compiler technology; architecturally K10's SSE support looks really good on paper. It seems to me that we just haven't seen code that was generated properly to take full advantage of it. And HT really doesn't buy anything if you can keep your active cores sufficiently fed that they have fewer stalls.

Yes, the i7 is one heck of a chip, it also carries one hell of a premium price tag as well with needing to have a new MB, expensive ram, etc. HT as you said, only can show any gain if the core's are not under full load. Under full load performance may actually slightly decrease.

Now onto SSE. Intels SSE does have a advantage in it's hardware. The intels cores have 3 128-bit units with 2 of them being symmetric vs the old K8's 2 64-bit units. With Phenom SSE performance did see an increase though with them now finally being able to execute two 128-bit generic SSE ops and 1 SSE MOV per cycle. Still the intels do carry a bit of an advantage here with it's 3 128-bit units.

Compilers do make a difference as well but greater gains can actually be had by optimizing the code for a specific processor that compilers may try to do but don't always succeed at. GMP for example when optimized with AMD specific patches abolutely will trounce on a intel with simular hand optimization (to be fair though how the i7 would respond isn't really known yet). Matlab, SuperPi, Prime95 is what intel guys love to show off for intel's "mathematical superiority" but all of those carry very intel specific optimizations or genaric optimizations without the same level of optimization done for the AMD specific CPU's.

[d2kx]

I did some first PTS tests with the Phenom II X4 940 today.

http://global.phoronix-test-suite.com/?k=author&u=d2kx-dragon

Nexuiz, OpenSSL and 7zip. Nexuiz in different settings (drivers, kernels).

What I found out so far: performance is awesome, CnQ does not slow down, 790GX/SB750 work fine with Linux although I used a HD 4850 1GB.

System was Debian/sid.

[Kano]

nexuiz can not use 4 cores, so a intel dual core with 2.4 ghz beats it (amd quad running at 3.0 ghz!), at 3.2 ghz even 33% - that bench scales really good from 9*266 to 8*400. Or maybe fglrx is just too slow ;)

http://global.phoronix-test-suite.com/?k=profile&u=kano-14926-3248-8692

[highlandsun]

Some basic bandwidth and FP number crunching tests I ran on Opteron 2384 and Xeon 5450

http://www.amdzone.com/phpbb3/viewtopic.php?f=52&t=135912&p=151532#p151532

In aggregate, 8 cores fully loaded, the Opteron still wins, but the Xeons win most of the single-thread runs. There's something really weird at the 2Kx2K size where the Xeon performance drops. Perhaps a bug in the FFTW library.

Also an interesting note - the code compiled with gcc -march=core2 runs slower on Opteron than code compiled for -march=amdfam10. Kind of expected. But the code compiled for -march=amdfam10 ran *faster* on the Xeon than the code compiled for -march=core2. gcc 4.3 is doing something weird...

[deanjo]

Some basic bandwidth and FP number crunching tests I ran on Opteron 2384 and Xeon 5450

http://www.amdzone.com/phpbb3/viewtopic.php?f=52&t=135912&p=151532#p151532

In aggregate, 8 cores fully loaded, the Opteron still wins, but the Xeons win most of the single-thread runs. There's something really weird at the 2Kx2K size where the Xeon performance drops. Perhaps a bug in the FFTW library.

Also an interesting note - the code compiled with gcc -march=core2 runs slower on Opteron than code compiled for -march=amdfam10. Kind of expected. But the code compiled for -march=amdfam10 ran *faster* on the Xeon than the code compiled for -march=core2. gcc 4.3 is doing something weird...


GCC's intel optimization has never been all that great. GCC on amd's on the other hand enjoy great architecture optimizations in GCC / nasm / yasm/ etc. Not surprising as intel wishes for people to purchase their compilers instead.