AMD Phenom SuperPi Performance

[hat]

My dual core Athlon did better at stock settings (2600MHz).

 

[DRDNA]

i hate amd fanbois

Hate is an awfully Strong word for being a fan of something ,try having compassion for us lower performing brothers

 

[theonetruewill]

My notebook T2500 Core Duo @ 2Ghz kicks out 1M SuperpiMOD1.5 at 30.75s at fsb166.

Now I'm sure if we all think back in history just a few short years ago, we can remember what happened when Intel introduced the first NetBurst chips vs. the pentium III chips at the time.

Of course at that time, AMD was kicking Intel's but with the Athlons/Opterons - AMD owned the server market, putting the Xeons to shame.

Now it appears that AMD may be walking down "netburst" avenue just like "almighty" Intel did not so long ago.

Unfortunately, AMD's pockets may be too shallow to recover - This is a truly worriesome condition. The last thing we need is no competition for Intel. On the other hand why don't we all visit:

http://www.tilera.com

and see who the true power house of the future will be.....hmmmmm.....are they on NASDAQ yet??

Regards,
jtleon

More worryingly, clock for clock this dwindles behind my aging Pentium M @1.86Ghz; let alone when overclocked to 2.2Ghz:shadedshu However as said the checksum doesn't work.

 

[[I.R.A]_FBi]

Hate is an awfully Strong word for being a fan of something ,try having compassion for us lower performing brothers

loathe?

 

[niko084]

loathe?

Same thing....

I find fan boys annoying all together, especially when they are customers and they throw a fit because I told them to get a 8800gts 640mb over a 1950pro for elite gaming 6 months ago and they didn't listen....

Now they say its "MY" fault their video is slower than they wanted....

Yes because I control ATI.

 

[Dippyskoodlez]

More worryingly, clock for clock this dwindles behind my aging Pentium M @1.86Ghz; let alone when overclocked to 2.2Ghz:shadedshu However as said the checksum doesn't work.

Sub 2Ghz the A64 has always performed VERY sub par.

Breaking 2.5Ghz, the A64 starts to scale much more effeciently.

Performance is not linear.

 

[Ketxxx]

Ditto, any A64 below 2.5GHz sucks hard. Go to 2.8GHz or faster, and the suckers start to scream.

 

[Random Murderer]

However as said the checksum doesn't work.

yes, yes it does.

 

[Basard]

boycott superpi

that makes a lot of sense, somehow a dual core 3800+ works faster according to techpowerup's benchmarks.... im sure amd would really spend years making a processor slower than their previous generation.

if i got my hands on a phenom proc, then ran superpi and got a god awful score like that, i think i would just run a DIFFERENT benchmark and post the results of that as well. if the program doesnt work well with multiple cores then WHY USE IT TO TEST MULTIPLE CORE SYSTEMS?? i dont get it, run a real benchmark...

not even worth the use of grammar.... YAWN

 

[mas0n]

for all we know he has the memory clocked at 200MHz with stupid timings. I don't trust a Super Pi result as a reference of true performance without also seeing the memory tab of CPU-Z.

 

[Random Murderer]

for all we know he has the memory clocked at 200MHz with stupid timings. I don't trust a Super Pi result as a reference of true performance without also seeing the memory tab of CPU-Z.

obviously you've never heard of coolaler.

 

[mas0n]

obviously you've never heard of coolaler.

Obviously you've never heard of skepticism. I don't care if God herself hands me half-assed results and verification, they are still half-assed.

 

[kwchang007]

Err my T7200 at 2 ghz on a 166 mhz fsb gets 27 secs. I know people have been saying it doesn't scale linearly.....but let's say core 2 does....it's still a big amount of scaling to get it up to core 2 efficiency. Take a 6000+ to a e6750....no competition. Another thing is...it's most probable that Phenom won't clock as well as Core 2 because Phenom has a 11 stage pipline vs Core 2's 13 stage pipline.

 

[hat]

Reminds me of previous generations Athlons pwning the P4s lol

 

[PVTCaboose1337]

This is so sad...

 

[hat]

On the other hand SuperPI is a single threaded app, you would need something else to measure it's true performance. Core 2 processors are still better though as shown by kwchang.

 

[kwchang007]

On the other hand SuperPI is a single threaded app, you would need something else to measure it's true performance. Core 2 processors are still better though as shown by kwchang.

Single threaded...yup. But from what I understand Core 2 has also beat it in Pov-ray...but I don't believe it's as big of a margin of victory. The big thing with native quad-core is the reduced latency, not bandwidth. So it'll be interesting to see how K10 performs, what I'm thinking based on these benchmarks (I'm not included those 3d mark numbers because they just seem outlandish) K10 will lag behind in most areas except for bandwidth and floating point operations, two strongs of K8. I also predict K10 to rule in the multi socket server sector. Also....if AMD completely switches to 65 nm.....I believe their price to performance ratio will come close to...if not beat Intel's. Oh and overclocking....probably will scale to ~3.5 ghz on 65nm.....but these are just my predictions.

 

[Ketxxx]

Punch for punch I'm going to call stalemate between Intel and AMD, for now.

 

[yogurt_21]

bout what I expected for super pi and all other non multithreaded apps. the strogn point of this cpu isn't going to be single core operation, it's gogin to be multi core operation. (as is already evident with the athlon x2's gaining more performance for the secong core that intel does.) so sure super pi ( a bench made eons ago lol) isn't the best thing to run with the phenoms, so what? it's not like you sit down after work, turn on your comp and run super pi for 3 hours before dinner. lol

 

[InfDamarvel]

I wonder why anyone thinks this is real, AMD woudln't release such a slow processor. Its either a flaw in super pi or its a fake. No way it would go back to Pentium D performance.

 

[Mussels]

it is only a 2GHz chip. While intels may beat them at current clocks, we dont know what release clocks will be - they could throw them out at 3.4Ghz, the 2GHz one could be a 35W low heat model etc etc.

Yeah, its slow - THIS one is slow. Doesnt mean the other 10 models will be slow too.

I wonder why anyone thinks this is real, AMD woudln't release such a slow processor. Its either a flaw in super pi or its a fake. No way it would go back to Pentium D performance.

Pentium D's (prescotts) needed 3.6-4.0Ghz to achieve this speed. Doing that at almost half the clocks is not the 'same speed' - i think the flaw lies within you, not superpi.

 

[driver66]

I love how people call this real http://www.techpowerup.com/?38494



And the same people call this thread fake? :shadedshu

 

[Mussels]

I love how people call this real http://www.techpowerup.com/?38494



And the same people call this thread fake? :shadedshu

Nevermind that its a different CPU, with a 50% higher clock speed. noooo, the 2Ghz model just isnt cool now. AMD should only release 4Ghz models, cause the AMD fans say so.

 

[affinity0]

Wow, my 2.6 X2 runs superpi 1M in 34s....that cant be right...

 

[Wile E]

I don't buy these benches, nor do I buy into 3DMark benches. I'll wait for them to lift the NDA, before I make a judgment.