Phenom/Barcelona Reviews/Previews and Pre-release Discussion

[SLee]

Found this at techreport. . .Oblivion bench run at medium settings at 800x600, so about as cpu dependant as you can get for that game,

http://www.techreport.com/articles.x/12616/3

TechReport uses their own FRAPs run-through plus they only use a single 7900GTX. The difference between a E6300 and a X6800 is only 21% in Tech Report vs 41% on Anandtech's test so the tests are not equivalent.

I don't think either of us will convince each other so I'll just make my prediction that a 2GHz Phenom will be as fast as a E6320 on Oblivion and HL2. Your prediction is it will match a E6420 and we'll see who's right in a couple of months.

 

[gone_fishin]

In an article,
http://www.crn.com/hardware/201805044
"The Test Center was provided with an engineering sample server equipped with dual Barcelona CPUs and 16Gbytes of RAM"

and used PassMark's benchtest to obtain a cpumark score of 4693 points.

Going to Passmark's site, you can see an updated table,
http://www.cpubenchmark.net/cpu_lookup.php?cpu=[Dual+CPU]+Intel+Xeon+E5345+@+2.33GHz

Including Xeon 5355 dual cpu (8 core it means) that get 6207 cpumarks.

So AMD = 4693 vs Intel's 6207.
And the highest end Intel 5365 is not on the charts yet.

 

[savageseb]

In an article,
http://www.crn.com/hardware/201805044
"The Test Center was provided with an engineering sample server equipped with dual Barcelona CPUs and 16Gbytes of RAM"

and used PassMark's benchtest to obtain a cpumark score of 4693 points.

Going to Passmark's site, you can see an updated table,
http://www.cpubenchmark.net/cpu_lookup.php?cpu=[Dual+CPU]+Intel+Xeon+E5345+@+2.33GHz

Including Xeon 5355 dual cpu (8 core it means) that get 6207 cpumarks.

So AMD = 4693 vs Intel's 6207.
And the highest end Intel 5365 is not on the charts yet.

you missed the part with the Dual cores 8218 2core *4 processor( 8 core total) scored 5028


and wait

omg!

there are no 4core * 2processor setups from AMD in the passmark charts, HOW COULD THAT BE!!!!!!!! lmao

Opteron 2350 is not shown in the charts.

not withstanding , that the Opteron 2350 is at launch clock speed of 2.0ghz, and amd is ramping to 2.5ghz by the end of the year. with native Quad Scaling way better we´ve yet to see the performance gains it will providenn and passmark has an incomplete chart...


btw has anyone found any benchmarking of an OC´d Opteron 23** series?

 

[hUMANbEATbOX]

not withstanding , that the Opteron 2350 is at launch clock speed of 2.0ghz, and amd is HOPING TO ramp to 2.5ghz by the end of the year.

There I fixed it for you. You shouldn't talk in absolutes, it is not a guarantee that AMD will be shipping 2.5ghz parts by the end of the year.

Also, when the 3ghz phenoms systems starting getting "demo'd", many took that to mean that AMD has fixed the leakage etc and that a 3ghz part would be imminent. ... Where is 3ghz on the roadmap now?

 

[gone_fishin]

The team from the article either did not submit the result or if you read Passmarks qualifications, no setup gets included in the chart unless there are at least 2 of the same processor systems submitted.
That said the article is a published piece and therefore the numbers have been published, like it or not.

 

[Shell]

You shouldn't talk in absolutes, it is not a guarantee that AMD will be shipping 2.5ghz parts by the end of the year.

I was told there will be 2.4 and 2.6 chips out this year.

 

[hUMANbEATbOX]

I was told there will be 2.4 and 2.6 chips out this year.

You were told by who? Doc Brown?

Seriously, the best laid plans of mice and men often fall asunder.

 

[ryanmartini]

Like Ive been saying for a long time now. Until I have a PHENOM in my DFI abusing the hell out of pc8500 ddr2 running 3dmark on my 8800gts, I could care less about computer enthusiasts arguing about Barcelona's performance in the server market.

Its performance is spectacular in edges of the market that buy up new servers such as virtualization and simulations. It was reported that the B1 stepping could clock to 3ghz. I cant wait to see the Phenom with a good chipset.

Just because Intel beats out AMD in some areas does not mean we do not have a fighter here. This is a great development to set the stepping methods in upcoming Bulldozer etc on a native quad platform.

AMD enthusiasts will be rewarded for their patience.

 

[Kuroimaho]

18 months for this then another 1.5 years for bulldozer.

 

[MadMan007]

There I fixed it for you. You shouldn't talk in absolutes, it is not a guarantee that AMD will be shipping 2.5ghz parts by the end of the year.

Also, when the 3ghz phenoms systems starting getting "demo'd", many took that to mean that AMD has fixed the leakage etc and that a 3ghz part would be imminent. ... Where is 3ghz on the roadmap now?

Normally I would agree with you on not guessing but one of the sites, I think Anand, had 2.5GHz samples in hand for testing. They were production chips not just one-off cherry picked ES's I believe, That's a pretty good sign they will in fact appear for real.

As for greater than 100% scaling that someone mentioned it shouldn't be possible theoretically especially if the bottleneck is systematic outside the core. Maaybe if increased core speed that includes faster cache as well as faster mem controller and the cores are that bandwidth-limited. One would hope that cache would be designed to overcome this but an on-die mem controller that increases speed can do funny things. Got to think about that one a little more...

 

[hUMANbEATbOX]

Normally I would agree with you on not guessing but one of the sites, I think Anand, had 2.5GHz samples in hand for testing. They were production chips not just one-off cherry picked ES's I believe, That's a pretty good sign they will in fact appear for real.

Yes and there were 3ghz 'samples' as well. Why doesn't Anand (or ANYONE) have one of those to test?

Samples mean nothing. Those weren't production chips, if they were, AMD would be shipping them now. They could be production stepping but that doesn't mean that they are getting good enough yields to launch them now or by the end of the year.

 

[MadMan007]

What real websites did reviews which included 3GHz samples?

 

[Kuroimaho]

None, just AMD insiders were able to touch that rig and the infamous inquirer to run the 30K+ 3DMark. Can't wait to see proof of that, so come on AMD give us the 3G phenom.

 

[hUMANbEATbOX]

What real websites did reviews which included 3GHz samples?

That's exactly my point! There were 3ghz machines demo'd by AMD at least 3 times about a month or so ago. Kuroimaho was at one! They ran no benches, but claimed that the computers were at 3ghz. Now, how come they are sending out 2.5ghz samples, but no 3ghz samples? Oh, and the 3ghz samples were Phenom and had tri-fire going on.

I'm not trying to rain on any parades, I hope they can ramp quickly myself, but don't count your chickens before they hatch.

 

[MadMan007]

Ok, feel free to cast some doubt upon the 3GHz chips but 2.5GHz looks like a pretty solid bet and that's all I'd mentioned and was the point of my question - 3GHz is 2.5GHz seems

 

[Mr.Guvernment]

are we getting into this "native" "not native" crap again.. i thought that died out with dual cores and crap...

 

[Kuroimaho]

No, AMD couldn't make quads so far so they were busy claiming the inferiority of the Intel chips, personally I do not want this to be forgotten yet, would be really funny if those inferior chips come out on top...

 

[ryboto]

That's exactly my point! There were 3ghz machines demo'd by AMD at least 3 times about a month or so ago. Kuroimaho was at one! They ran no benches, but claimed that the computers were at 3ghz. Now, how come they are sending out 2.5ghz samples, but no 3ghz samples? Oh, and the 3ghz samples were Phenom and had tri-fire going on.

I'm not trying to rain on any parades, I hope they can ramp quickly myself, but don't count your chickens before they hatch.

Aren't the Agena cores different than Barcelona? Also, as far as the info from the Xtremesystems thread, the stepping that's shipping now is BA, a bios patched B1 revision. So, the B2, which clock to 2.5, weren't produced in volume until recently, which means they wont be ready for a little while. At least, that's the puzzle that I've pieced together.

 

[Kuroimaho]

Didn't most of this info come from a guy who claimed he had one before launch and it's like 20%+ faster than C2D, then after the anand benches he said we just misinterpret him ?

I am certainly not buying any of these in the near future, as things look new steppings will continue to roll out.

 

[e6600]

You were told by who? Doc Brown?

Seriously, the best laid plans of mice and men often fall asunder.

george is representative of intel
lennie is amd

lennie makes several mistakes throughout the story, all building up to his eventual downfall
george is forced to put down lennie, because lennie has not learnt from his mistakes and probably never will