Super PI 1M scores...

i posted it in cpu cause its a cpu test...

ok now that that has been said, my super pi 1m score is 38 sec... is that good? check sig for system specs

conroe 2.4ghz = ~20seconds aparently

my P4c 2.6c @ 3000 gets 45 but after a reinstall it went up to like 48? odd

I get 43 on my XP Barton 2600+ @ 2.3ghz

I got 30.734 last night on my overclocked AMD Athlon 64 3700+ San Diego Core CPU, which is at 2.69Ghz.

makes my celeron D 2.4@3000 look slightly better - it gets bout that, maybe a little less

41sec with a 2800+ sempron s754@2.3Ghz and 1Gig of DDR @ 192Mhz.

1m 01sec at stock 2800+ speed (1600Mhz) and RAM @ 134Mhz.

Is this test useful for anything? just curious that's all since so many people like it.

38.250 seconds:

Pi fast (version 4.3) time 2.39 seconds.
Pi fast (version 4.3) 10,000,000 digits time 38.63 seconds.

P4EE 3.40GHZ Gallatin 478 pin FGA, 512KB L2 cache, 2MB L3 cache
130 nm technology, 1066mhz FSB.
Intel D875PBZ motherboard, 2GB 2-2-2-5 400mhz ddr ram.

34.953 seconds:

Pi fast (version 4.3) time 2.48 seconds.
Pi fast (version 4.3) 10,000,000 digits time 40.22 seconds
(yes, the 3.73ghz cpu was slower on the longer calculation).

P4EE 3.73GHZ Prescot 775 pin LGA, 2MB L2 cache, 0MB L3 cache
90 nm technology, 1066mhz FSB.
Intel D975XBX motherboard 2GB 4-4-4-12 667mhz ddr2 ram.

Do a web search for super-pi. It's been updated to so show milliseconds and includes a checksum.

Pifast is a lot faster, you can test it with 10,000,00 digits in not much longer than super-pi takes to do 1,000,000 digits.

superpi is a benchmark, and like any other is useful for comparing relative speeds between systems to a standard reference. superpi focuses on floating point computations...


Your statement is totally busted. What do you think the X2 4200+ compares to jajig's system 2800+ sempron? Both CUPs get identical 41 sec. Do they have the same ability? That's ridiculous if you say such.

Intel P4 3.0Ghz 530J, 1MB L2cache, oc'ed to 3.27Ghz.
Mushkin Enhanced PC3500 DDR433 512MB RAM 2-3-3-6, overclocked to DDR440, 2T 2-3-3-6, 2.75 volts.
I got 42 seconds. I need to find the more accurate version.

superpi is a benchmark, and like any other is useful for comparing relative speeds between systems to a standard reference. superpi focuses on floating point computations...
no good conclusions should be formed from a solitary reference mark; but a wide variety of benchmarks focusing on specific tasks that push the limits of performance in a particular direction you are interested in will give you an idea of the general throughput that pc has in the genre of processes you are interested in.

SuperPI is no longer a CPU test, but rather a "How much Cache you got?" Test, simple as that. Having a SuperPI of 5s doesn't mean your CPU is "pwnage", it means you probably have 1GB of L2 Cache.

~~Mad Mod Mike, pimpin' the world 1 rig at a time

superpi is a benchmark, and like any other is useful for comparing relative speeds between systems to a standard reference. superpi focuses on floating point computations...
no good conclusions should be formed from a solitary reference mark; but a wide variety of benchmarks focusing on specific tasks that push the limits of performance in a particular direction you are interested in will give you an idea of the general throughput that pc has in the genre of processes you are interested in.

SuperPI is no longer a CPU test, but rather a "How much Cache you got?" Test, simple as that. Having a SuperPI of 5s doesn't mean your CPU is "pwnage", it means you probably have 1GB of L2 Cache.

~~Mad Mod Mike, pimpin' the world 1 rig at a time

Conroe's scores did not result only from an increased L2 cache. Heck, if my P2 had 4MB L2, it would probably be only 2s faster, at about 40s.
A: Conroe had all around faster scores with other benchmarks.
B: SuperPi takes much more than 4MB to calculate, it takes 10MB at least.

4MB did not make the difference. Notice how the mobile processors got 25seconds? They did not have 4MB L2.... they had only 2MB. Same as a 6xx series P4. They were much faster than the P4's. I think you've gone off the deep end with your cache theory.

superpi is a benchmark, and like any other is useful for comparing relative speeds between systems to a standard reference. superpi focuses on floating point computations...
no good conclusions should be formed from a solitary reference mark; but a wide variety of benchmarks focusing on specific tasks that push the limits of performance in a particular direction you are interested in will give you an idea of the general throughput that pc has in the genre of processes you are interested in.

SuperPI is no longer a CPU test, but rather a "How much Cache you got?" Test, simple as that. Having a SuperPI of 5s doesn't mean your CPU is "pwnage", it means you probably have 1GB of L2 Cache.

~~Mad Mod Mike, pimpin' the world 1 rig at a time

Conroe's scores did not result only from an increased L2 cache. Heck, if my P2 had 4MB L2, it would probably be only 2s faster, at about 40s.
A: Conroe had all around faster scores with other benchmarks.
B: SuperPi takes much more than 4MB to calculate, it takes 10MB at least.

4MB did not make the difference. Notice how the mobile processors got 25seconds? They did not have 4MB L2.... they had only 2MB. Same as a 6xx series P4. They were much faster than the P4's. I think you've gone off the deep end with your cache theory.

A: Which can be explained by the Cache & DDR2
B: SuperPI will fit nicely in 4MB of Cache

NextPart:

The mobile processors get 25s because they use an architecture that is not sh!t, like a P4 is sh!t, you could give it 10MB of cache and it wouldn't help NetBurst.

You're just an idiot fanboy who won't accept that Conroe is 75% hype and 25% performance.

~~Mad Mod Mike, pimpin' the world 1 rig at a time

This is the most idiotic post I've ever seen you make. The Conroe is like those mobile processors YOU just talked about. Conroe uses an architecture that is not "sh!t"

A: if it's just the cache, then my P4 should be way faster with 4MB L2.
B: if Conroe is faster in all the other benchmarks too, (even if it is just because of large L2), then it is faster in real world performance.
C: are you going to deny that the FPS that Conroe got in gaming is real world performance?

superpi is a benchmark, and like any other is useful for comparing relative speeds between systems to a standard reference. superpi focuses on floating point computations...
no good conclusions should be formed from a solitary reference mark; but a wide variety of benchmarks focusing on specific tasks that push the limits of performance in a particular direction you are interested in will give you an idea of the general throughput that pc has in the genre of processes you are interested in.

SuperPI is no longer a CPU test, but rather a "How much Cache you got?" Test, simple as that. Having a SuperPI of 5s doesn't mean your CPU is "pwnage", it means you probably have 1GB of L2 Cache.

~~Mad Mod Mike, pimpin' the world 1 rig at a time

Conroe's scores did not result only from an increased L2 cache. Heck, if my P2 had 4MB L2, it would probably be only 2s faster, at about 40s.
A: Conroe had all around faster scores with other benchmarks.
B: SuperPi takes much more than 4MB to calculate, it takes 10MB at least.

4MB did not make the difference. Notice how the mobile processors got 25seconds? They did not have 4MB L2.... they had only 2MB. Same as a 6xx series P4. They were much faster than the P4's. I think you've gone off the deep end with your cache theory.

A: Which can be explained by the Cache & DDR2
B: SuperPI will fit nicely in 4MB of Cache

NextPart:

The mobile processors get 25s because they use an architecture that is not sh!t, like a P4 is sh!t, you could give it 10MB of cache and it wouldn't help NetBurst.

You're just an idiot fanboy who won't accept that Conroe is 75% hype and 25% performance.

~~Mad Mod Mike, pimpin' the world 1 rig at a time

This is the most idiotic post I've ever seen you make. The Conroe is like those mobile processors YOU just talked about. Conroe uses an architecture that is not "sh!t"

A: if it's just the cache, then my P4 should be way faster with 4MB L2.
B: if Conroe is faster in all the other benchmarks too, (even if it is just because of large L2), then it is faster in real world performance.
C: are you going to deny that the FPS that Conroe got in gaming is real world performance?

You're not understanding fanboy....it, it's not registering up in your head...

I said P4 is SH!T, and P-M Arch is NOT SH!T

The only benchmarks I've seen of gaming were from Intel @ IDF and Anand, which were machines given to them by Intel, if there are others, I'd like a link to view them myself

I deny that Intel provided real tests, since even that "FX-60" @ "2.8GHz" was outperformed by a stock FX-60 in other tests, oh I forgot, for some reason, you can't compare those results to other tests, I guess it's cuz those other ones are fake and only Intel's are real, darn!

~~Mad Mod Mike, pimpin' the world 1 rig at a time

i think your score is about right. i got 38swith venice 3200+ @ 2.4ghz

superpi is a benchmark, and like any other is useful for comparing relative speeds between systems to a standard reference. superpi focuses on floating point computations...
no good conclusions should be formed from a solitary reference mark; but a wide variety of benchmarks focusing on specific tasks that push the limits of performance in a particular direction you are interested in will give you an idea of the general throughput that pc has in the genre of processes you are interested in.

SuperPI is no longer a CPU test, but rather a "How much Cache you got?" Test, simple as that. Having a SuperPI of 5s doesn't mean your CPU is "pwnage", it means you probably have 1GB of L2 Cache.

~~Mad Mod Mike, pimpin' the world 1 rig at a time

Conroe's scores did not result only from an increased L2 cache. Heck, if my P2 had 4MB L2, it would probably be only 2s faster, at about 40s.
A: Conroe had all around faster scores with other benchmarks.
B: SuperPi takes much more than 4MB to calculate, it takes 10MB at least.

4MB did not make the difference. Notice how the mobile processors got 25seconds? They did not have 4MB L2.... they had only 2MB. Same as a 6xx series P4. They were much faster than the P4's. I think you've gone off the deep end with your cache theory.

A: Which can be explained by the Cache & DDR2
B: SuperPI will fit nicely in 4MB of Cache

NextPart:

The mobile processors get 25s because they use an architecture that is not sh!t, like a P4 is sh!t, you could give it 10MB of cache and it wouldn't help NetBurst.

You're just an idiot fanboy who won't accept that Conroe is 75% hype and 25% performance.

~~Mad Mod Mike, pimpin' the world 1 rig at a time

This is the most idiotic post I've ever seen you make. The Conroe is like those mobile processors YOU just talked about. Conroe uses an architecture that is not "sh!t"

A: if it's just the cache, then my P4 should be way faster with 4MB L2.
B: if Conroe is faster in all the other benchmarks too, (even if it is just because of large L2), then it is faster in real world performance.
C: are you going to deny that the FPS that Conroe got in gaming is real world performance?

C: Are you going to deny that once Conroe is released and is all hype like the Intel dual-cores were and AMD kicks its ass? I mean, it's Intel your trusting, since when did their word satisfy people when the chips were released? Definately not with the dual-cores, because they sucked @$$.

Intel did not perform or report the benchmarks, btw...

And yes, if Conroe does not perform I have no problem admitting that I was bamboozooled... but frankly, all reports with credibility point to the fact that Conroe does perform decent, better than any current desktop CPU's, and I am pretty sure Intel's pricing is going to be roughly on target... as well as I am sure too that it will be available in quantity at launch and on and on...
Even if AMD manages to pull a rabbit out of their hat, and performance jump for AM2 boards is 20%, they will still lose the price wars, since they are illustrating quite clearly that they can not drop prices to compete with Conroe's product lines.

What about you? All bets/threats go two ways... When the cards are laid out... Are you going to admit your beloved and dear AMD has failed you again? Punishment will be to purchase the slower FX-62 for the low "at AMD's cost" of $1,100 upwards, (what a deal, since they are factory overclocked and no room left for you to overclock due to repeated and random BSOD's and freezing: and they run mucho hotter and on older architecture) while I get a Conroe for around $300 or so that equals or exceeds performance and has tons of headroom for overclocking all while running cool and quiet!!!
Muhahahahahahahha!!!

Yeah tell them how it is.Fanboy comments are not welcomed.Atleast by me i hate fanboys.

Now Mike, you should know this is not true. SuperPI is a tunable benchmark, you can change the parameter (number of digits of Pi calculated) from less than 1M to up to 32M. Even if you believe that the 1M test fits in a 4MB cache (which it does not), the 32M benchmark tests much deeper sections of the memory hierarchy, including the system RAM and hence (gasp), the FSB.

Now even on the 32M test Conroe is completely kicking ass. Here is the checksummed, verified score for the 32M test of 21 Minutes, which is way way faster than even the highest overclocked, supercooled AMD systems, and this is on a 2.4GHz stock Conroe:

SuperPi 32M 2.4GHz Conroe verified score

Now all this means is that Conroe is really fast and efficient at doing Floating point operations. It is just one benchmark and for the purposes of this thread we can just conveniently ignore all of the other released benchmarks where Conroe wins by a similar margin, it's okay. I suppose you could choose another straw to grasp at and say that Floating Point no longer matters in the real world.

Best of luck!

Hmm....interesting....NOT

Okay fanboy, let me analyze this....

I see Conroe benchmarks for gaming....done by Intel & AnandIntel, both on the SAME machines.....using "Modified ATI Drivers to 'recognize' the CPU" (yea, totally dude).....Not allowed to peek into "Device Manager" (hmmm...wonder what's in there )......than we get VictorWang's, his prove Conroe is good at SuperPI....hmmm.....sooo???

Than we see Clovertown compared in 64-bit....and it gets wooped by a 2800+....but wait, clovertown is NGMA? WTF? hmm...this is interesting....hmm....than we see all 8 cores, and yet, we don't see 8x the performance...but rather 3.7x....hmmm...interesting again....

Than we see all the fanboys (JumpingJack, You, ltcommander) saying "It's just an ES, it can't be taken as final" but we see AM2 ES's and you all say "that's the final product, it sucks" hmm...so who's the worse fanboy? All the AMD fanboys who give many links to benchmarks done NOT by Intel that prove NGMA < K8, yet you still don't realize it? Hmm....this is quite intresting....

Cache Thrashing - Not possible? Think again, even if the CPU manages it....guess what that means.... CPU Overhead! No way, dude, totally man!

I wait for your fanboy post to reply against mine, I should get a good laugh or 2 out of it. I'll start ahead of time....HaHaH-(awaits your post to finish laughing)

~~Mad Mod Mike, pimpin' the world 1 rig at a time

Hmm...you're an Idiot? yes, you are.

Immature? Hey, He started it! (lol)

Top once again? I can't remember...oh wait, ya, the P3...30 years and 1 good CPU, OH SNAP!

Punch in the face? hmm....now who's immature?

~~Mad Mod Mike, pimpin' the world 1 rig at a time

superpi is a benchmark, and like any other is useful for comparing relative speeds between systems to a standard reference. superpi focuses on floating point computations...
no good conclusions should be formed from a solitary reference mark; but a wide variety of benchmarks focusing on specific tasks that push the limits of performance in a particular direction you are interested in will give you an idea of the general throughput that pc has in the genre of processes you are interested in.

SuperPI is no longer a CPU test, but rather a "How much Cache you got?" Test, simple as that. Having a SuperPI of 5s doesn't mean your CPU is "pwnage", it means you probably have 1GB of L2 Cache.

~~Mad Mod Mike, pimpin' the world 1 rig at a time

Why isnt it a CPU test it calculates pi which last I checked was done via math. As per the cache deal would't your cache thrashing theory cause Conore to flop at the test?

i may be immature, since u know... im still a teenager, but that doesnt change the fact that your still a blatent AMD fanboy that needs to have his fingers chopped off and voicebox cut (so u cant use the speech translator) u dont belong in these forums because ur a troll. all u do is put down p4 because the prescott was bad. member northwood? that pwned as well. i guess u forgot how AMD had heat problems. even in 64bit era, ever see a clawhammer core? i have one. 50C idle with 4 case fans and a thermaltake blue orb. yes i applied the paste right. its just a hot chip, amd bartons were hot as hell too. if ur gonna hate a company based on past experiences or present problems then u should hate amd due to there hot cpu's. im gonna personally PM u and laugh in ur face when conroe comes out and it rips shit all over amd2. conroe is gonan be way ahead of amd2 due to the lowered power consumption and being on 65nm. so even if amd2 is faster than conroe, they can just bump the clocks up like crazy. new p4's on 65nm are known to hit 4.5 on air EASILY, try it with vapochill and u can hit 6ghz. easily faster than any amd2 will be on 90nm. all this, and im an amd fan. but its people like you that make me want to switch to some noname processor brand... maybe ill make 1.

Can we get Off CPU L2 Chip? Like even if the L1 might be a bit off sync with the seperate L2 we could just increase the speed to recover the non-syncrinousness so now INTEL can start providing 256Mb and 512Mb or even 1gig L2 Chips that would be rokka$$!!

Wow....you are now the spokes-boy for the Intel DumbF*ck fanboys. You are surely a hillarious person mr. Rich, I think it's common, everybody I've ever met who's been named "Rich" (excluding 1 person, sorry, not u) has been a complete moron, and you surely do not disappoint!

Conroe != Faster, and surely is != Cooler....Quieter???? WTF is that? I didn't know CPU's made noises, surely it is the CPU HSF, but hmm.....that's interesting, hearing a CPU....hmmm...dillusional? Yes, I think so.

BSOD's? Sir, have you been in a closet? Must I post the 50 links of AMD's @ 3GHz-4GHz for you? I will if I must.

Sorry Rich, again, your fanboyism has overshrowed you and you are once again DEAD WRONG.

~~Mad Mod Mike, pimpin' the world 1 rig at a time


8O 8O 8O
mike,,, I thought you knew me a bit better... sorry for that slip, but knowing me the sarcasm was piled on high for that one it should have been obvious for someone who reads some of my posts... geek humor if you will, granted not very good humor, but humor nonetheless...

but I see no reason to call me names and point out my stupidity, especially seeing how obvious it is.

Lighten up my friend, and laugh a little (in the positive way that is... with me and not at me ) There is enough dumbledorfs stumbling around in our twisted consciousness that seize every opportunity to laugh at the mistakes and eagerly await the chance to ridicule the trails of others at their expense, that we tend to forget that we are all in the same boat, and ultimately are here to help one another and do our part to make this place a little better than it was before. You mileage may vary!

Blah Blah Blah is all I hear, along with attempts to patronize and condescend me, which make me laugh more. You've said in the past an AMD @ or above 3GHz is hard, yet AMD has Opty's @ 3GHz w/ 95w TDP's and I have over 50 links of AMD CPU's from 3GHz-4.1GHz. Hmm....Not to mention PERSONAL OC's of 3GHz.....hmmm.....

~~Mad Mod Mike, pimpin' the world 1 rig at a time

Now Mike, you should know this is not true. SuperPI is a tunable benchmark, you can change the parameter (number of digits of Pi calculated) from less than 1M to up to 32M. Even if you believe that the 1M test fits in a 4MB cache (which it does not), the 32M benchmark tests much deeper sections of the memory hierarchy, including the system RAM and hence (gasp), the FSB.

Now even on the 32M test Conroe is completely kicking ass. Here is the checksummed, verified score for the 32M test of 21 Minutes, which is way way faster than even the highest overclocked, supercooled AMD systems, and this is on a 2.4GHz stock Conroe:

SuperPi 32M 2.4GHz Conroe verified score

Now all this means is that Conroe is really fast and efficient at doing Floating point operations. It is just one benchmark and for the purposes of this thread we can just conveniently ignore all of the other released benchmarks where Conroe wins by a similar margin, it's okay. I suppose you could choose another straw to grasp at and say that Floating Point no longer matters in the real world.

Best of luck!

Hmm....interesting....NOT

Okay fanboy, let me analyze this....

I see Conroe benchmarks for gaming....done by Intel & AnandIntel, both on the SAME machines.....using "Modified ATI Drivers to 'recognize' the CPU" (yea, totally dude).....Not allowed to peek into "Device Manager" (hmmm...wonder what's in there )......than we get VictorWang's, his prove Conroe is good at SuperPI....hmmm.....sooo???

Than we see Clovertown compared in 64-bit....and it gets wooped by a 2800+....but wait, clovertown is NGMA? WTF? hmm...this is interesting....hmm....than we see all 8 cores, and yet, we don't see 8x the performance...but rather 3.7x....hmmm...interesting again....

Than we see all the fanboys (JumpingJack, You, ltcommander) saying "It's just an ES, it can't be taken as final" but we see AM2 ES's and you all say "that's the final product, it sucks" hmm...so who's the worse fanboy? All the AMD fanboys who give many links to benchmarks done NOT by Intel that prove NGMA < K8, yet you still don't realize it? Hmm....this is quite intresting....

Cache Thrashing - Not possible? Think again, even if the CPU manages it....guess what that means.... CPU Overhead! No way, dude, totally man!

I wait for your fanboy post to reply against mine, I should get a good laugh or 2 out of it. I'll start ahead of time....HaHaH-(awaits your post to finish laughing)

~~Mad Mod Mike, pimpin' the world 1 rig at a time

I still don't see what your issue is with the benchmarks done by Intel and Anandtech, it's not like they can lie about the performance that’s false advertisement as well this isn't Apple we are talking about.

The clovertown debate has been done and over with. It's been concluded the software used was not what the shak fellow claimed it was. As well doctoring to the pictures has been confirmed as well.

On a side not the performance was 4.7 not 3.7 and to make matters more skewed K8, Netburst, and Core based machines score very close to each other regardless, which in turn invalidates the software's benchmark feature.

As well it was Anandtech that claimed that the AM2 performance doesn't appear to be increasing by any sizeable amount by the time the machines hit the market. If you have a problem with their conclusions, perhaps you should take a stroll to Anandtech and try voicing your thoughts there.

CPU overhead? You will have to elaborate further as well I think its time you make the point clear as to why you believe "cache thrashing" will cause an issue. Which by all information provided by Intel insists that it’s not plausible with the dynamic allocation of the cache subsystem or as Intel coins it "Intel Advanced Smart Cache".