Sign in with
Sign up | Sign in
Your question

Conroe SuperPi Scores

Last response: in CPUs
Share
a b à CPUs
April 4, 2006 9:52:03 PM

Stumbled across some SuperPi scores for a conroe, scored 21.4 seconds and that was for the 2.4ghz mode, comparable to a Pentium M at 4ghz, a P4 Prescott at 7.2ghz, and an A64 at 4ghz.

http://www.xtremesystems.org/forums/showthread.php?t=95...

Just thinkin - what performance would we see at 2.93ghz or 3.33ghz?

And as for the AMD fanboys there im already expecting you to post replys saying how fake the benchmarks are and whatever - its expected from such fanboy low life :D  - enjoy.

Oh an did you year - AMD's 65nm Could come as late as next year - that means a limit of ~3ghz for AMD's 90nm design, a sad time for AMD if it does take that long.

*waits for AMD to drop prices to get a bargain AMD*

More about : conroe superpi scores

April 4, 2006 9:58:54 PM

Wow, now thats impressive. Interesting that cpuz says it has SSE4.
April 4, 2006 9:59:13 PM

Quote:
Stumbled across some SuperPi scores for a conroe, scored 21.4 seconds and that was for the 2.4ghz mode, comparable to a Pentium M at 4ghz, a P4 Prescott at 7.2ghz, and an A64 at 4ghz.

http://www.xtremesystems.org/forums/showthread.php?t=95...

Just thinkin - what performance would we see at 2.93ghz or 3.33ghz?

And as for the AMD fanboys there im already expecting you to post replys saying how fake the benchmarks are and whatever - its expected from such fanboy low life :D  - enjoy.

Oh an did you year - AMD's 65nm Could come as late as next year - that means a limit of ~3ghz for AMD's 90nm design, a sad time for AMD if it does take that long.

*waits for AMD to drop prices to get a bargain AMD*


It's interesting to see how this "Conroe" does 75%+ better in SuperPI, but in the Intel Benchmarks, it only did up to 20%. Hmm...

Apache: You are a stupid fanboy. Period.

~~Mad Mod Mike, pimpin' the world 1 rig at a time
Related resources
April 4, 2006 10:07:17 PM

Ok Ok Ok...

"AMD is the best in AMD world" and "Intel is the best in Intel world"....

Phew... Now it is off my chest, just joking :lol:  :lol: 
can all of the expert explain what is Super Pi test and how it test the CPU/System?

Thanks.

Chime.
April 4, 2006 10:08:23 PM

Quote:
Stumbled across some SuperPi scores for a conroe, scored 21.4 seconds and that was for the 2.4ghz mode, comparable to a Pentium M at 4ghz, a P4 Prescott at 7.2ghz, and an A64 at 4ghz.

http://www.xtremesystems.org/forums/showthread.php?t=95...

Just thinkin - what performance would we see at 2.93ghz or 3.33ghz?

And as for the AMD fanboys there im already expecting you to post replys saying how fake the benchmarks are and whatever - its expected from such fanboy low life :D  - enjoy.

Oh an did you year - AMD's 65nm Could come as late as next year - that means a limit of ~3ghz for AMD's 90nm design, a sad time for AMD if it does take that long.

*waits for AMD to drop prices to get a bargain AMD*




Dude you sound like a little kid. AMD is not going anywhere and probably WON'T drop their prices in response to chip that is not available. I hope Intel/AMD doesn't look like Cyrix/Intel anymore but this is ridiculous. AMD is the second CPU producer, not some fly by night operation run by drunks.
:twisted:
Just like Chartered magically popped 65nm out of their asses, AMD will not just "go away." 20% on AM2 X2/FX with process tweaks and 1066 DDR2 with low latencies. OCZs newest DDR2 will soup the mem ctrlr like putting a jet engine on a turbo prop plane. Don't say I didn't tell you. months ago.
April 4, 2006 10:19:52 PM

I don't recall any superpi benchmarks.
April 4, 2006 10:23:09 PM

Quote:
I don't recall any superpi benchmarks.


Going by physical performance, if it takes a 4GHz AMD64 to get 21s and a 7GHz P4 to get 21s, and supposedly a 2.4GHz "Conroe" gets 21s, and it's only a glorified P3/PM, you're telling me to believe that? Okay, maybe I am a f*cking fanboy idiot, but no way in hell can tell me that is true. If I saw a new AMD K8L @ 2.4GHz getting that, I'd call BS as well, there's no physical way that @ 2.4GHz got 21s in SuperPI 1M. If that's true, that would mean the 2.66GHz @ IDF would be like a 8GHz P4, but how come it didn't beat the sh!t out of the @$$ of the "FX-62"? Exactly, that is the biggest BS of XtremeSystems ever seen, even die-hard fanboys have to call BS on this.

~~Mad Mod Mike, pimpin' the world 1 rig at a time
April 4, 2006 10:26:31 PM

I think I might be tired of all this crap for a mere couple of seconds, framerates...etc...

-My di*k is bigger than yours!
-But mine is thicker!

-I'm f*** 2 girls in 1' 23".
-I overclocked my amtel cojones and now do that in 1' 20".

But this will never end... :roll:
April 4, 2006 10:32:48 PM

Well I've never used superpi but I do know a little about it. Dothan apparently tops the 1m scores because of its 2mb cache, conroe does have 4mb which can be used by one core. The checksum matches as well, not that I know if that makes any sort of difference.
April 4, 2006 10:35:50 PM

Quote:
Well I've never used superpi but I do know a little about it. Dothan apparently tops the 1m scores because of its 2mb cache, conroe does have 4mb which can be used by one core. The checksum matches as well, not that I know if that makes any sort of difference.


I don't doubt SOMETHING got 21s, but it sure as hell wasn't a 2.4GHz Conroe, I don't care if it had 1GB of L2 Cache.

~~Mad Mod Mike, pimpin' the world 1 rig at a time
April 4, 2006 10:36:56 PM

The price drop on AMD was already on the way in Q3. The drop is to fund R&D.
April 4, 2006 10:49:03 PM

Quote:
I'd call BS as well, there's no physical way that @ 2.4GHz got 21s in SuperPI 1M. If that's true, that would mean the 2.66GHz @ IDF would be like a 8GHz P4, but how come it didn't beat the sh!t out of the @$$ of the "FX-62"?
Glad we agree; I'm sure the Conroe will be fast, but there's no way a 2.4GHz conroe can do 1M Super PI in 21 seconds.
April 4, 2006 10:50:14 PM

Quote:
Stumbled across some SuperPi scores for a conroe, scored 21.4 seconds and that was for the 2.4ghz mode, comparable to a Pentium M at 4ghz, a P4 Prescott at 7.2ghz, and an A64 at 4ghz.

http://www.xtremesystems.org/forums/showthread.php?t=95...

Just thinkin - what performance would we see at 2.93ghz or 3.33ghz?

And as for the AMD fanboys there im already expecting you to post replys saying how fake the benchmarks are and whatever - its expected from such fanboy low life :D  - enjoy.

Oh an did you year - AMD's 65nm Could come as late as next year - that means a limit of ~3ghz for AMD's 90nm design, a sad time for AMD if it does take that long.

*waits for AMD to drop prices to get a bargain AMD*


It's interesting to see how this "Conroe" does 75%+ better in SuperPI, but in the Intel Benchmarks, it only did up to 20%. Hmm...

Apache: You are a stupid fanboy. Period.

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

I see your point but, Intel didn't use super pi in their benchmark testing. What caught my eye the most is the voltage.

We'll just have to wait and see when they come out just how well they do. Anything is better at this point than Intels current technology you gotta at least admit that.
April 4, 2006 10:53:18 PM

Quote:
I'd call BS as well, there's no physical way that @ 2.4GHz got 21s in SuperPI 1M. If that's true, that would mean the 2.66GHz @ IDF would be like a 8GHz P4, but how come it didn't beat the sh!t out of the @$$ of the "FX-62"?
Glad we agree; I'm sure the Conroe will be fast, but there's no way a 2.4GHz conroe can do 1M Super PI in 21 seconds.

Yes, Conroe will definately be a good chip, but that fake SuperPI crap from Fanboy sites like KiddySystems.org is starting flaming that is unnecessary and down right BS.

@Lumi: I realize that, but 2 things got my attention:

1) If that score is due to the 4MB L2 Cache, than "Conroe" itself ISN'T the good chip, it's just due to it's cache

2) A revamped P3/PM can beat a K8 clock-for-clock by that much? No way.

I am no way saying Conroe will suck, but it's not THAT good, and there's no way anybody can contest it.

~~Mad Mod Mike, pimpin' the world 1 rig at a time
April 4, 2006 10:59:13 PM

Quote:
1) If that score is due to the 4MB L2 Cache, than "Conroe" itself ISN'T the good chip, it's just due to it's cache


Well dothan wins the 1m benchmark because of its cache, once you go higher the K8 is meant to win. Since there are no benchmarksdoing >1m we can't say.
April 4, 2006 11:02:08 PM

Quote:
1) If that score is due to the 4MB L2 Cache, than "Conroe" itself ISN'T the good chip, it's just due to it's cache


Well dothan wins the 1m benchmark because of its cache, once you go higher the K8 is meant to win. Since there are no benchmarksdoing >1m we can't say.

So if we go by the Theory of Cache, once AMD uses Z-RAM/L3 Cache, it will skyrocket above Conroe? For some reason, that theory bounces off of me....

~~Mad Mod Mike, pimpin' the world 1 rig at a time
April 4, 2006 11:08:28 PM

Its possible, such chips don't exist and details are still thin with Z-RAM and the L3 cache.
April 4, 2006 11:17:37 PM

Quote:
Its possible, such chips don't exist and details are still thin with Z-RAM and the L3 cache.


I just can't see Cache being the reason it did so well, I could see maybe it getting 28-29s due to Cache and new architecture if it was @ 2.66GHz, but 2.4GHz and 21s? Last time I checked, Hell wasn't blue.

~~Mad Mod Mike, pimpin' the world 1 rig at a time
April 4, 2006 11:29:48 PM

That Conroe Engineering Sample has sold on eBay... so it might not be too long before people start reporting thier benchmarks!!! :p 
April 4, 2006 11:31:55 PM

Quote:
That Conroe Engineering Sample has sold on eBay... so it might not be too long before people start reporting thier benchmarks!!! :p 


A box with the word "Conroe" on it doesn't mean it's real Rich, sorry.

~~Mad Mod Mike, pimpin' the world 1 rig at a time
April 4, 2006 11:34:48 PM

Quote:
Last time I checked, Hell wasn't blue.


Hell isn't blue, but you'll be using an intel conroe processor to play all your games in hell.... If you're lucky :wink:

I have a feeling though that you'll get a p4 prescott w/ netburst technology. :lol: 
Yea... you'd hate that a lot more. Plus it's hell, its gotta be hot!!!
April 4, 2006 11:35:58 PM

Quote:
Last time I checked, Hell wasn't blue.


Hell isn't blue, but you'll be using an intel conroe processor to play all your games in hell.... If you're lucky :wink:

I have a feeling your gonna get a p4 prescott w/ netburst technology in hell. :lol: 
Yea... you'd hate that a lot more.

Hmm.......

~~Mad Mod Mike, pimpin' the world 1 rig at a time
April 4, 2006 11:40:19 PM

Quote:
What caught my eye the most is the voltage


That is interesting... CPU-Z might have been configured wrong from the ini file, and was reading the 3.3v.

But that was allot of picture editing done on that cap.
April 5, 2006 12:04:46 AM

and what score gets an intel dual-core like a 955 or a 965

i mean, intel has always good benchmark scores.

and while it comes to a real application, it gets beaten...

75% is too much, i agree, if it does it, it's a simple optimisation

if 4mb cache gets 21 second, any pentium D would get it


BUT i get this :

P4C800-E dlx Intel PM S479 1833 @2695 25
SLI-DR Expert Amd Opteron dual 2200 @3200 27


intel mobile architecture was clearly optimized for these bench
as the P-M beats the a64 dual core on a lower clock with a single core...
and seriously, P-M aren't that much powerfull. conroe is just a dual core version of a boosted P-M... nothing more. so another line that says


so... conroe DOES 21s at 1M superpi

and ...



YOU CAN'T TRUST INTEL IN BENCHMARKS
April 5, 2006 12:15:08 AM

Actually a Pentium D is 2x2MB cache, so each core can only use 2MB of cache. On the conroe one core can use all of the available 4MB of cache.
April 5, 2006 12:16:23 AM

Quote:
Actually a Pentium D is 2x2MB cache, so each core can only use 2MB of cache. On the conroe one core can use all of the available 4MB of cache.


I've got a question, do the cores in Conroe communicate via the FSB or over the unified L2 Cache?

~~Mad Mod Mike, pimpin' the world 1 rig at a time
April 5, 2006 12:19:30 AM

shouldnt cpu-z not recognize a conroe core yet?... wouldnt it just say unidentified or somthin... thats what strikes me as odd since cpu-z prob hasnt been updated for conroe
April 5, 2006 12:21:21 AM

They can do L1 and L2 core to core.

Quote:
shouldnt cpu-z not recognize a conroe core yet?


1.33 has some support for conroe.
April 5, 2006 12:24:50 AM

Quote:
They can do L1 and L2 core to core.

shouldnt cpu-z not recognize a conroe core yet?


1.33 has some support for conroe.

So the 2 cores are Coherent across the Level 1 AND Level 2 Cache?

~~Mad Mod Mike, pimpin' the world 1 rig at a time
April 5, 2006 12:31:24 AM

BUT i get this :

P4C800-E dlx Intel PM S479 1833 @2695 25
SLI-DR Expert Amd Opteron dual 2200 @3200 27
P4C800E dlx Intel PIVE Prescott 3000 @4705 00:00:29
Epox 8RDA+ Amd Duron 1800 @2457 00:00:47
M320 Rev 1.0 Intel 33 339:08:55

intel mobile architecture was clearly optimized for these bench
as the P-M beats the a64 dual core on a lower clock with a single core...
and seriously, P-M aren't that much powerfull. conroe is just a dual core version of a boosted P-M... nothing more. so another line that says


so... conroe DOES 21s at 1M superpi

and ...



YOU CAN'T TRUST INTEL IN BENCHMARKS
April 5, 2006 12:33:54 AM

What do you mean?

Quote:
intel mobile architecture was clearly optimized for these bench


Not really, the assloads of cache helps it out.

Quote:
conroe is just a dual core version of a boosted P-M... nothing more.


mmhmm.
April 5, 2006 12:35:53 AM

Quote:
What do you mean?


In the Pentium D, the 2 Processor Cores communicate to each other over the Front Side Bus, on an Athlon X2, the 2 Processor Cores communicate to each other over the CrossBar, how does Conroe communicate the 2 Cores? Over the Front Side Bus? If so, than Conroe will still get beat in Multi-Threaded applications by a K8, regardless of the architecture.

~~Mad Mod Mike, pimpin' the world 1 rig at a time
April 5, 2006 12:39:22 AM

It can transfer directly between the L1 caches. It also uses the L2.
April 5, 2006 12:41:40 AM

Quote:
It can transfer directly between the L1 caches. It also uses the L2.


Accordign to the Conroe Block Diagram, the L1 Cache is connected directly to the L2 Cache as well, but the L1 Cache in each core DOES NOT connect directly to the cores, so either the 2 cores communicate via the L2 Cache or via the FSB. I am thinking it's not possible for them to communicate via the L2 Cache, but I may be wrong, anybody know?

~~Mad Mod Mike, pimpin' the world 1 rig at a time
April 5, 2006 12:41:47 AM

what, only 21.4 seconds, I'm going faster on my 90nm single core bugest processor :lol: 
April 5, 2006 12:46:34 AM

Quote:
what, only 21.4 seconds, I'm going faster on my 90nm single core bugest processor :lol: 


you're not talking about the same test
April 5, 2006 12:48:07 AM

Quote:
one chip of 4mb shared between the 2 cores i would say according to intel roadmaps showing Pentium D with 2x2mb 2x1mb and conroe as 2mb 4mb

http://images.anandtech.com/reviews/tradeshows/IDF/2005...


Okay than it looks like the only problem is still the Coherency in Woodcrest/Dempsey in Multi-CPU Configurations.

~~Mad Mod Mike, pimpin' the world 1 rig at a time
April 5, 2006 1:02:24 AM

can we then assume the tests intel has choosen was in relation with conroe cache memory boost


like runing a game in full resolution full quality AND

DUAL SCREENED
April 5, 2006 1:05:38 AM

Quote:
can we then assume the tests intel has choosen was in relation with conroe cache memory boost


Nope, they tested quake 4 with SMP.
April 5, 2006 1:10:13 AM

Quote:
can we then assume the tests intel has choosen was in relation with conroe cache memory boost


Nope, they tested quake 4 with SMP.

Once you get into High-Resolutions, SMP Gaming is useless, which is why I don't get why everybody is saying "Wait til Dual-Core Games for greater improvement", it's proven w/ CoD2 and Quake 4, that Multi-Core CPU's isn't as good as Multi-Core (SLI/CF) GPU's.

~~Mad Mod Mike, pimpin' the world 1 rig at a time
April 5, 2006 1:15:42 AM

conroe still have 4mb cache while using a single core...
so using benchmarks which CAN use 4mb cache will increase it's performance while apps which are not cache limited won't get that gain...
April 5, 2006 1:18:55 AM

Quote:
Once you get into High-Resolutions, SMP Gaming is useless, which is why I don't get why everybody is saying "Wait til Dual-Core Games for greater improvement", it's proven w/ CoD2 and Quake 4, that Multi-Core CPU's isn't as good as Multi-Core (SLI/CF) GPU's.


CoD2 and Quake 4 have at best average implementations of SMP.
April 5, 2006 1:27:41 AM

since smp WON'T limit the cpu cache to 2mb on conroe, smp doesn't change anything...

cache is still to 4mb and they just had to use cache-intensive bench to gain 25% over 1mb max cache cpu's

so intel's advance is still and has always been cache memory...

2mb cache over a K8 would increase the gain in the same benchs
April 5, 2006 1:31:09 AM

Quote:
since smp WON'T limit the cpu cache to 2mb on conroe, smp doesn't change anything...


Well yeah since it has 4mb but each core would use ~2mb each.

Quote:
cache is still to 4mb and they just had to use cache-intensive bench to gain 25% over 1mb max cache cpu's


What are you on?

Quote:
so intel's advance is still and has always been cache memory...


Clearly. :roll:
April 5, 2006 1:31:42 AM

Quote:
I've got a question, do the cores in Conroe communicate via the FSB or over the unified L2 Cache?

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


Could you loose that stupid remark, please? The world aint your 2gig whore.

There is only one L2 cache and it is shared by both cores. One core could use all of it, or just a small part of it. Some of it can be turnd off when it's not needed to save power. And each core can read any of it!
April 5, 2006 1:54:56 AM

Quote:
I've got a question, do the cores in Conroe communicate via the FSB or over the unified L2 Cache?

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


Could you loose that stupid remark, please?

There is only one L2 cache and it is shared by both cores. One core could use all of it, or just a small part of it. Some of it can be turnd off when it's not needed to save power. And each core can read any of it!

You obviously don't know what the hell I'm talking about, so you're being ignored...

@Vern: That Ageia PhysX is pointless and worthless, and doesn't improve performance for anything.

~~Mad Mod Mike, pimpin' the world 1 rig at a time
April 5, 2006 1:56:05 AM

how about let's just wait until Conroe's out, and some "real" performance tests can be done? ever since this "Conroe" myth came out earlier this year, everyone is going out crazy, throwing meaningless insults at each other.

can't everyone just calm down and wait until Conroe actually arrives
April 5, 2006 2:01:42 AM

the physics card doesnt do anything in games now because they werent programed for it... future games that are heavy on physics, which a lot promise to be, will be improved a ton by the card...
!