3.7E vs 3.7C becouse we really wanna know who wins right?

deathstar13 · Jun 18, 2004

pls this is a freindly comparison and lets try and keep it that way for the sake of learning and expanding our views,and over come the ones we are fed.

with that said on to the fun!
its been rather known and i agree the C chip at or below 3.2-3.4ghz tends to outrun the E chips,but i hope to show what happens at higher mhz.
tom tried to do this also but failed miserably imo.
http://www6.tomshardware.com/cpu/20040517/index.html
nothing was run in benchmarks we use to show performance.his is more scientific,im more of a dirt and rocks kinda basic performance minded person.

i also would like to mabey see how this scales on up higher to 4ghz if we can get that high with another C.

so onto the show.
heres what benches ive run and open to others:
sandra arith,multi media
3dmark2k2
super pi

i cant run 3dmark atm to compare unless you have an original g-force256 32mb card lol
my new one is comeing soon so mabey then ill have a 9800pro to work with.

all benches were done at as close to 3.7ghz as i can.i wanna keep this a str8 cpu comparision and not have anything that can interfere with scores,say memory and such.

Jimbob7 · Jun 19, 2004

I could only get my 3.0C as high as 3.6

, so that will have to do for the moment.

In any case the scores are rather interesting, The C edition still just pulls out the win even though its @ 3.6. Although i think as you start getting 3.9+ Overclocks, its going to get close if not even. With the new D0 Stepping coming out soon 4G on air could become alot less rare, making the presscott the overall faster CPU. A 4G+ overclock with a presscott should be the ' Northwood Killer ' hehe, just gotta sit back and wait.

CandymanCan · Jun 19, 2004

wow the 3.6c crushes the 3.7E.

If you want i can post some AMD scores @ 2.62ghz ?

Jimbob7 · Jun 19, 2004

I think deathstar13 just wants to compare Presscott vs Northwood within this thread.

CandymanCan · Jun 19, 2004

I wasj ust trying to be nice, would be good to see results from abroad.

Jimbob7 · Jun 19, 2004

Yeh, i understand, but i think this thread would just become an Intel VS Amd war.

CandymanCan · Jun 19, 2004

Yea ur right some people just cant handle it i guess, thats sad..

Anyway its interesting to see that a 3.6c beats the 3.7E in sandra... I wonder why u got a higher score.

Jimbob7 · Jun 19, 2004

The northwood has shorter pipelines to the Presscott, but the presscott is going to be able to reach alot higher speeds, so it will be faster. The northwood seems to be maxing out as to how far intel can push it i think.

Captain Newbie · Jun 19, 2004

[BBE]Jimbob said:
The northwood has shorter pipelines to the Presscott, but the presscott is going to be able to reach alot higher speeds, so it will be faster. The northwood seems to be maxing out as to how far intel can push it i think.

Stuff that Intel doesn't tell you:

A longer pipeline demands a higher clockrate. Unfortunately for Prescott, Intel has hit a wall at 3.4. A longer pipeline also demands better branch target prediction as a longer pipeline must stall for a longer amount of time on a missed branch. Prescott implements a larger BTB but shows no better prediction than Northwood.

A Prescott should always be theoretically slower than a Northwood at equivalent clock rates. This discounts HTT, cache sizes and latencies, ... and speaks of pure pipeline architecture.

It's like a tall person and a short person running a mile. The tall person has longer legs, and therefore takes fewer but longer strides (shorter pipeline, lower frequency like Northwood and by extension AMD's procecssors) to run the mile. The short person (Prescott) takes more strides, but each is shorter--sometimes drastically shorter--than that of the tall person's strides.

deathstar13 · Jun 19, 2004

wow,i must say im suprised at the scores myself.
[BBE]Jimbob ty for posting those.how about a pcmark2k2 run?

CandymanCan-i only want to compare E vs C here not amd VS intel.i would just not be interested in that arguement atm.i can handle it but it isnt what i want to accomplish.
also a thread like that would end up being locked fast.if not on its own but id ask my thread be locked.ive seen too much bad attitudes from people on that side.sad as thats where i came from myself.

Captain Newbie- your analogy sounds right and seems to fit the results.

now i know batboy can do 3.8ghz on his C mabey some higher mhz C's want to see when or if these chips start running even mhz for mhz? im starting to wonder if they do?

Jimbob7 · Jun 19, 2004

I would have run Pc mark 2002 @ 3.6 but this chip is really messed, just get sudden reboots about 30 seconds into pc mark. Im suprised it managed to get through sandra to be honest.

I think the main advantage of the presscott is going to be the fact that it will be able to hit high speeds, i really do think the northwood core is maxing out. So even if the presscott doesn't become equal clock for clock with the northwood, it doesn't matter, because it will achive higherspeeds, whereby it will be faster than the northwood.

deathstar13 · Jun 19, 2004

[BBE]Jimbob said:
I would have run Pc mark 2002 @ 3.6 but this chip is really messed, just get sudden reboots about 30 seconds into pc mark. Im suprised it managed to get through sandra to be honest.

eekk sorry to here that i see in your sig also now.

im seeing alot of people saying the sandra benchmarks are are heavily biased in the C cores favour.
i dont want to dismiss the benchmark but would rather see how true the biased rumor is.
i dont know if pcmark2k2 would see this nor be proof enough.
but anyhow at least now i do see how and why people say the things they do about the Emhz vs Cmhz now.

Jimbob7 · Jun 19, 2004

If i remember correct, my 2.4c @ 3.5 got around 8k in pc2002 (standard test) although don't quote me on that.

deathstar13 · Jun 19, 2004

[BBE]Jimbob said:
If i remember correct, my 2.4c @ 3.5 got around 8k in pc2002 (standard test) although don't quote me on that.

well i do tend to belive you as even with my bad memory just some things stick in you head.

hhmmmm i cant think of any other benches that just do cpu's and will isolate it.
even super pi can be biased with memory timings.

i want to stay objective,but these sandra scores seem odd.mabey thats why we dont see any comparisons besides toms as there may not be a decent way to compare these?

i dunno id like more feedback on this.

batboy · Jun 19, 2004

I can't hardly read those screenshots of yours, Deathstar. Looks like my 3.0C at 3.7 gig wins some and loses some. The Prescott does good in Super Pi, Sandra multi-media, and PCmark2002 memory scores.

Sandra CPU ALU:
3.0/3.7=11403

Sandra CPU FPU:
3.0/3.7=3212/7170

PCmark2002 CPU:
3.0/3.7=9194

PCmark2002 memory:
3.0/3.7=11599

PCmark2002 HDD:
3.0/3.7=1250

Super Pi:
3.0/3.7=38s

jpm804 · Jun 19, 2004

Here are some of the benchamrks I ran:

Gautam · Jun 19, 2004

Guys, it's no surprise that the Prescott is losing in Sandra. It's such a focused test that the Prescott's longer pipeline doesn't allow it to stand a chance. Try out SuperPI, PIFast, 3DMark, UT, etc., and you will see a very different story. PCMark also has a series of focused tests, so I wouldn't put much faith in those either. Focused tests like these are useful for seeing what sort of architecture the processors have, not what the implications of them are.

batboy · Jun 19, 2004

We are running Super Pi (or at least Deathstar and I are). 3Dmark and UT are worthless unless everyone has the same vid card. PCmark2002 is actually a pretty good all purpose benchmark.

JPM, you need to run those benchies at 3.7 gig for this to be a valid comparison.

Is it just me or do other people have trouble reading those screenshots?

deathstar13 · Jun 19, 2004

batboy said:
We are running Super Pi (or at least Deathstar and I are). 3Dmark and UT are worthless unless everyone has the same vid card. PCmark2002 is actually a pretty good all purpose benchmark.

JPM, you need to run those benchies at 3.7 gig for this to be a valid comparison.

Is it just me or do other people have trouble reading those screenshots?

yeh i need more practice getting the size kb down before sizing.

sandra arith: alu: 9869 fpu:4264/7696
multimedia: int:26159 floating:34729
pcmark2k2: cpu 7787

super pi is heavily dependant on memory bandwith and timings unfortunatly.but we can still see some things from it.

Gautam- i believe alot in your remarks about the accuracy of the tests showing performance that can be judged against other cpu's

tomorow my daughter goes home so i need to get off here and spend time with her,but im gonna do some research on some things we can use to compare better.i gotta find something better my pressy is looking baaaaadd lol

Pitbull2k · Jun 21, 2004

250fsb 3500mhz

255fsb 3570mhz (running now but cant prime

)

3.7E vs 3.7C becouse we really wanna know who wins right?

FSB FRIEK

Attachments

Member

Disabled

Member

Disabled

Member

Disabled

Member

Senior Django-loving Member

FSB FRIEK

Member

FSB FRIEK

Member

FSB FRIEK

Senior Moment

Member

Senior Benchmark Addict

Senior Moment

FSB FRIEK

Registered

Similar threads