Conroe @ 2.66Ghz vs FX-57 @ 3.6Ghz: 3DMark05 CPU Test

freecableguy ***xtremesystems.org exclusive information***

Benchmark: 3DMark05 CPU Tests - single threaded application
AMD FX-57 @ 3.6Ghz vs. Intel "Conroe" ES @ 2.66Ghz

CPU: AMD 0516 FX-57 @ 3.6Ghz (257x14) @ 1.7v
Memory: 2x512MB PC-3200 Mushkin BH-5 @ 2-2-2-5, 257Mhz 1:1
Motherboard: DFI LanParty Venus NF4 SLI-DR
Cooling: Custom 'Chilly1' single-stage autocascade
CPU Temps: -30C idle, -15C full load while benching
OS: WinXP Pro SP2

CPU Test 1: 4.0 fps (compared to Conroe: -35%)
CPU Test 2: 7.4 fps (compared to Conroe: -13.5%)

FX-57 @ 3.6Ghz score: 8146 (compared to Conroe: -23%)

http://members.cox.net/kjboughton/FXG_FX8146.jpg


CPU: Intel "Conroe" 2.13Ghz ES @ 2.66Ghz (333x8) @ 1.15v
Memory: 2x512MB PC2-4200 Mushkin @ 3-2-2-4, 333Mhz 1:1
Motherboard: Intel D975XBX "Bad Axe"
Cooling: Stock Intel aluminum heatsink, ambient ~22C
CPU Temps: 35C idle, 41C full load while benching (according to EVEREST)
OS: WinXP Pro SP2

CPU Test 1: 5.4 fps (increase over FX-57: +35%)
CPU Test 2: 8.4 fps (increase over FX-57: +13.5%)

Conroe @ 2.66Ghz score: 10038 (increase over FX-57: +23%)

http://members.cox.net/kjboughton/FCG_C10038.jpg


-FCG xs.org

Thanks to Absolute_0 for the following benchmark for comparison:

CPU: AMD Opteron 170 @ 2.66Ghz (266x10) @ 1.325v
Memory: 2x1GB ??? @ 3-4-3-8-1T, 266Mhz 1:1
Motherboard: ???
Cooling: water cooling
CPU Temps: not reported
OS: WinXP Pro SP2

CPU Test 1: 4.2 fps (compared to Conroe: -28.5%)
CPU Test 2: 7.4 fps (compared to Conroe: -13.5%)

Opteron 170 @ 2.66Ghz score: 8347 (compared to Conroe: -20%)

http://img241.imageshack.us/img241/8...parison2fu.jpg

really nice :clap:
me want one :D

No way... get out of town! :slobber:

What can i say, a very big
:clap: :clap: :clap: for FCG

Amazing results... just great. Very professional and well documented benchmark too, i like the blue and green color coding. Great Job Kris.

I know this was some work, but can we look forward to this kind of comparison for other benchmarks too? :D

Damn.... it really does look like clock for clock it is about 40% better. I seriously hope the haters don't give you are hard time about the graphics cards being different. Here is a preemptive: "For god's sake, this is the CPU test, the graphics cards don't matter!".

Cheers, FCG, keep it up!

very very nice comparison. I hope you will do more of this style in the future.

:toast: :toast: :toast: :toast:

Too bad you don't have AMD X2 to compare with, though ;)
From what we've seen on the ORB, X2 needs ~3.0GHz to match 5.4fps/8.4fps... That would give Conroe healthy 300-350Mhz as far as 3D benchmarking goes.

Very nice FGC, glad to see some user-end comparisons... like others have said, it would be great if you could do some other benches, but I know you did a lot of work for what you gave us!

Thanks man, looking forward to more results.

Quote:

---

Originally Posted by bachus_anonym

Too bad you don't have AMD X2 to compare with, though ;)
From what we've seen on the ORB, X2 needs ~3.0GHz to match 5.4fps/8.4fps...

---

Sounds about right, 05 benefits a decent bit from dual core. Still 2.66 Conroe @ stock volts and stock heatsink beating 3 Ghz X2? sounds good to me. And beats 3.6 Ghz FX57!! lol, single core is out :p:

Quote:

---

Originally Posted by Absolute_0

Sounds about right, 05 benefits a decent bit from dual core. Still 2.66 Conroe @ stock volts and stock heatsink beating 3 Ghz X2? sounds good to me. And beats 3.6 Ghz FX57!! lol, single core is out :p:

---

BTW it's a 2.13 conroe oc'd to 2.66... not quite stock, but it still makes me happy ;)

Quote:

---

Originally Posted by Burdman

BTW it's a 2.13 conroe oc'd to 2.66... not quite stock, but it still makes me happy ;)

---

Yes i know, but stock volts and stock heatsink... and the overclock is still restricted by the board! It's comparable to X2 @ 3 Ghz, i'd like to see that on stock volts with stock heatsink! ;)

Just my speculation, i'll try not to get too carried away. I just absolutely cannot wait to see what Conroe 3.33 Ghz does, and how far it clocks :toast:

look, i'm loving what conroe is showing, but putting a single core against a dual core isn't fair, single-threaded application or not, and there was no need to have seperate vid cards, as you could just use 1 card, and use the same in each setup, and if possible, use an opty dual-core, or an x2 with 2x1mb L2, and of course, run them both at same clock speed

what would also be nice is if you used 2x512mb for both, and find some tccd that can run 333

again, mucho appreciation for all your work, and again, i know conroe is powerful, but leave nothing for the flame-o's to feed on

edit: this would be the best possible scenario to show what cpu is better

3.6Ghz vs. 2.66Ghz isn't fair either....there is no fair.....let it go.

Lol awsome!!!!!!! :slobber:.You know what would be funny to see is a non FX cpu thats dual core and clocked to 2.6. I wonder how much faster that conroe will be HAHA! 60%? :P
Great work man!

Quote:

---

Originally Posted by freecableguy

3.6Ghz vs. 2.66Ghz is fair either....there is no fair.....let it go.

---

haha man i have been here since April 1st and I have seen that line pop up about million times....

Getting a HEADACHE ! haha Awesome scores..."off to make another thread in EOCF :P "

thanks for the comparison freecableguy, looks to have made up my mind about my next upgrade being conroe based :)

Great post FCG,

However, 3DMark05 does use multithreading in the CPU Tests. I am pretty sure vertex shaders are done with both processors, so you are really comparing Single core vs Dual Core. Conroe is definately an awesome chip though.

I just looked back at my own Dothan vs Yonah results, my previous Dothan @ 2.65GHz gets about 4700 CPUMarks, whereas my current Yonah setup gets 8874 CPUMarks. Sure, I may have different ram, mobo, etc..... but that big of a difference in my scores does not come from the new architecture alone, but rather because its dualcore and it does make use of it in the benchmark.

Another link showing Dualcore vs Singlecore 3Dmark05
http://www.xbitlabs.com/articles/cpu...x2-3800_5.html

Would be good to add Yonah and X2 @ ~3Ghz just for information :)

the difference is just huge :eek:

Alright i've run a comparison of my own

AMD Opteron 170 @ 2.66 Ghz
10x266 1:1 2gb ram @ 3-4-3-8-1T

3dmark05 CPU score = 8347 (approx. 17% slower than Conroe)
CPU test 1: 4.2 FPS
CPU test 2: 7.4 FPS

http://img241.imageshack.us/img241/8...parison2fu.jpg

Exactly my point...
This is a dual-core vs Single-Core comparison :o

Good comparison absolute.

So the result isn't THAT big a deal after all.. clock for clock, less than 20% on this test. Don't get me wrong, that's awesome, but it wasn't fair to test against a single core apparently :stick:

Well it's pretty obvius the cpu scores are higher because it works kinda like 3ds max rendering. Dual core will always have a big lead there. I'm more interested how much difference there is on the final score with same GFX and software etc etc.

Thx absolute and now at 3ghz ? :toast:

thanx absolute, that at least shows that the raw speed of fx-57 got beat by a much lower clocked(clock and fsb) opty 170

now, fcg, if u could do a run again @ 266x8, and absolute, you the same, and both run your cards at close as possible speeds, you would end the possibility of any flaming

again, i know conroe is faster, and one should hope so, lol, it is just a 'best case scenario' to squash the almost inevitable flaming

and to think, you're still on a board with very limited overclocking ability, or, at least, we hope is limited(wouldn't want that to be the best board to support conroe, lol)

Alright i've re-ran trying to match Conroe's speed, came up short

AMD Opteron 170 @ 3.10 Ghz
10x310 166divider 2gb ram @ 258 3-4-3-8-1T

3dmark05 CPU score = 9565 (approx. 4.7% slower than Conroe)
CPU test 1: 5.1 FPS
CPU test 2: 8.0 FPS

http://img285.imageshack.us/img285/5...arison20re.jpg

Got about 12% performance for going up 440 Mhz, to equal Conroe i estimate closer to 3.3 Ghz needed, or 3.2 Ghz and waay faster ram than i've got.

Later tonight i can try benching at 3.2+ Ghz, we'll see if this thing can take on Conroe ;)

Thanks for the work guys, great stuff.

Can't wait to see more!

Thx absolute. That made the comparison kinda fun.

Great work guys, definitely the type of comparisons that are fun to see :D

I just found an almost perfect match with AMD FX-60 @ 253x12=3036Mhz & x1900xtx 786/882 ---> 5.4fps / 8.3fps

So, we get ~350Mhz headstart for Conroe :) Replace FX-60 with Opteron 170 for a bit better $$$ comparo since there will be some of those that hit 3GHz... The good ones, but still good just for kicks :)

Speaking of fast, new WR in 3DMark01 "Lobby - High" which is incredibly CPU intensive. Previous record was about 270 fps with a P4 up at around 7.1Ghz+...

http://members.cox.net/kjboughton/FCG_01details.jpg


-FCG

Quote:

---

Originally Posted by Absolute_0

Alright i've re-ran trying to match Conroe's speed, came up short

AMD Opteron 170 @ 3.10 Ghz
10x310 166divider 2gb ram @ 258 3-4-3-8-1T

3dmark05 CPU score = 9565 (approx. 4.7% slower than Conroe)
CPU test 1: 5.1 FPS
CPU test 2: 8.0 FPS

http://img285.imageshack.us/img285/5...arison20re.jpg

Got about 12% performance for going up 440 Mhz, to equal Conroe i estimate closer to 3.3 Ghz needed, or 3.2 Ghz and waay faster ram than i've got.

Later tonight i can try benching at 3.2+ Ghz, we'll see if this thing can take on Conroe ;)

---

hmm absolute hv a look at my score. unfortunately i dont have a "proper" account of 3dmark05 or i would post it to ORB to clear any doubts.....

i guess at 3.1Ghz I would score around 10.5/11k? Correct me if I am wrong.

I dont know why you are scoring hardly 300 points over me w/ over 400 Mhz in clock speed. Are the dual core drivers installed properly? What about the ram timings? Also see if you have forced AA or AF in the drivers. Just my 2 cents....

All runs that all three of us did (FCG, absolute_0, and I) were just close everything down and run it.

We're not doubting your results, but they're way better than what even bachus found on the ORB....did you do anything special? Your CPU Test 2 lines up with all the results we've seen, but your CPU Test 1 is just WOW.

CPU tests in futuremark are not very reliable as i found out a while back when i started comparing my first dual core results and got some info from eva2000 which made me realise after further testing that one should not rely on those scores very much..............i am talking about the CPU scores not actual 3D scores.........


nice w.r. there FCG :)

any fx-60 comparison???

That was the old WR, CF at 700/800

http://fugger.netfirms.com/51k.JPG

Quote:

---

Originally Posted by Vapor

All runs that all three of us did (FCG, absolute_0, and I) were just close everything down and run it.

We're not doubting your results, but they're way better than what even bachus found on the ORB....did you do anything special? Your CPU Test 2 lines up with all the results we've seen, but your CPU Test 1 is just WOW.

---

is it possible due to the tight timings (3-4-3-0, hv a look at a64info shot, everything is as tight as possible @ 265 Mhz w/ Samsung UCCC) ?

how i ran the test:

- started 3dmark05
- closed explorer.exe down
- ran 3dmark05

:woot:

Quote:

---

Originally Posted by FUGGER

That was the old WR, CF at 700/800

http://fugger.netfirms.com/51k.JPG

---

Looks like Fugger and FCG are playing " anything you can do I can do better."

heh, nah if that were the case I would have bumped AM3 WR.

@ this speed I score over 140K with no OC on VC in AM3.

Dang Conroe really does kick azz :clap:

Lol Kris, I removed his post.

Quote:

---

Originally Posted by FUGGER

Lol Kris, I removed his post.

---

lol guys.... FCG replaces his SS with a nice custom made :banana::banana::banana::banana: pic and u remove the post....:woot:

great post/results. thanks for sharing all of this, especially with all of the updates to clear things up and show exactly where it stands in regards to other chips at various clocks.

Awesome!

Quote:

---

Originally Posted by bachus_anonym

I just found an almost perfect match with AMD FX-60 @ 253x12=3036Mhz & x1900xtx 786/882 ---> 5.4fps / 8.3fps

So, we get ~350Mhz headstart for Conroe :) Replace FX-60 with Opteron 170 for a bit better $$$ comparo since there will be some of those that hit 3GHz... The good ones, but still good just for kicks :)

---

I guess the ram is why it outscored me, i'm guessing the 253 is 2-2-2-5 of course

@ fcg's request (he asked me to recreate the score and post a ss similar to his):

AMD Opteron 165 @ 2.655Ghz @ 1.54Vcore

3DMark05 Score = 9407

http://img83.imageshack.us/img83/9025/recreated4au.jpg

well i gave it my best shot.... the cpu1 test does sway around a bit.... sometimes 4.8, sometimes 5.3. this is the best overall score i could get.

will try to get it further up when I return from university.

Nice Scores

Confusion :(

3.1 = 9565 and 2.66 = 9407 :confused:

Quote:

---

Originally Posted by kidoman

@ fcg's request (he asked me to recreate the score and post a ss similar to his):

AMD Opteron 165 @ 2.655Ghz @ 1.54Vcore

3DMark05 Score = 9407


well i gave it my best shot.... the cpu1 test does sway around a bit.... sometimes 4.8, sometimes 5.3. this is the best overall score i could get.

will try to get it further up when I return from university.

---

what ram timings?

And best overall score, how many runs? I did best of two, seemed fair

Quote:

---

Originally Posted by FUGGER

That was the old WR, CF at 700/800

http://fugger.netfirms.com/51k.JPG

---

Finally a couple of 2001 scores for me to pour over .. really really nice considering the air cooling :toast: :toast:

7Ghz LN2 Intel gets beaten by air cooled Conroe in lobby ... Intel sure made the right choices months / years back on changing direction.

Thanks for the info FUGGER and FreeCableGuy.

Regards

Andy

Conroe still has almost 2000 more points in CPU test over the opty 170 @2.66!
Nice to see a clock to clock comparison between these 2 CPU's!
Great guys, thanks for all info you provide!

Quote:

---

Originally Posted by Burdman

BTW it's a 2.13 conroe oc'd to 2.66... not quite stock, but it still makes me happy ;)

---

It's what a stock E6700 would be, so it is a fair comparison.

Quote:

---

Originally Posted by Durzel

It's what a stock E6700 would be, so it is a fair comparison.

---

Except for the FSB speed :)

Quote:

---

Originally Posted by Thorburn

Except for the FSB speed :)

---

Yep your right. Its just gona be silly when we find out how high a Conroe can overclock on air. I think way over 3Ghz ;)

Fun days await us :toast:

THX for the guyz reviewing and making the comparisons on there own rigs.:)

@ FCG did u take the WR on the HIGH LOBBY scene in 3DMark 01 ? Or did u miss it with 3 frames compared to the 7.1Ghz P4 :p:

Quote:

---

Originally Posted by Thorburn

Except for the FSB speed :)

---

True, true ;)

Quote:

---

Originally Posted by ORCBEAST

THX for the guyz reviewing and making the comparisons on there own rigs.:)

---

Couldn't agree more! :clap:

Quote:

---

Originally Posted by FUGGER

That was the old WR, CF at 700/800

http://fugger.netfirms.com/51k.JPG

---

Sweet jesus they are fast in 01:slobber: look at CL,LL all for that matter.....CRAZY...i love it:slobber: :D and only with 2700mhz .........wow!

why does the 2.6 ghz opty have close to the same score as the 3.6ghz fx?

Because the Opteron is dual core, the FX is single core

was just about to edit that

now i really want a conroe of my own :)

i can't wait to see amd compete with their di-shrink, consumers win, dimes win, rosetta wins, fanboys lose

Quote:

---

Originally Posted by VulgarHandle

i can't wait to see amd compete with their di-shrink, consumers win, dimes win, rosetta wins, fanboys lose

---

:toast:

Quote:

---

Originally Posted by mesyn191

*Repost from other thread*

Uhhh guys, PCmark05 is hardly a valid realworld test...

Its tends to favor Intel chips heavily. An example:

http://www.xbitlabs.com/articles/cpu...x2-3800_5.html

---

Lucky this thread is about them testing with 3D Mark 05 instead then ;)

My bad, even worse than pcmark tests though as 3dmark is just a glorifed video bench, but I'll delete my other post.

Nice Screens and work guys.
Conroe seems to be a good one.
Seems to be good to have 2x2mb L2 cache.
Because the FX and the Optys have only 1mb or 2x1mb L2 cache.

Conroe has 4MB L2 - not 2x2MB like it is with Preslers.

http://www.ocheaven.com/article/0509...0909-10-03.jpghttp://www.cnet.com.au/i/r/2005/desk...s7_440x330.jpg

Yep, Conroes cache design allows the amount of L2 cache available to each core to be varied, from a 2MB each split to one core taking control of all 4MB in intensive single threaded calculations.

now if they could do the same for the cores themselves...

Quote:

---

Originally Posted by VulgarHandle

now if they could do the same for the cores themselves...

---

Well technically thats what Hyperthreading did to a degree, allowing one cores resources to execute more than one thread if it had the capacity to do so....

Sorry Cooper saw it too late.
4Mb shared L2 Cache for the E6600 and 6700.
And the little ones below only 2Mb L2 Cache.
Just caught it in the Inet.Sorry

Nothing to sorry about. We all make mistakes and of course we can`t know everything :)

Sorry for OT

Just googled a little bit and found that.

http://www.computerbase.de/news/hard...rom_woodcrest/

For the Future an Woodcrest 5130 would be also very interesting. Price and FSB Speed 1333 and Cache 4MB.
Price around 320Dollars

Yonah @ 2.66GHz VS. Conroe @ 2.66GHz

Anyone to test out ?

thanks for the comparison

Quote:

---

Originally Posted by wa77ss

Yonah @ 2.66GHz VS. Conroe @ 2.66GHz

Anyone to test out ?

---

I put my money on Conroe. Any takers? :D

No contest. Everything has shown Conroe to have better IPC than Yonah ;)

and a real benchmark? one, at least

I think Conroe would beat a Yonah for sure .... but I would just like to see a percentage of performance increase like they did with the FX-57/Opteron 170

Quote:

---

Originally Posted by FUGGER

That was the old WR, CF at 700/800

http://fugger.netfirms.com/51k.JPG

---

OMG so Nice Charles!!!...missed this. Thats crazy air power!

Is it too hard to post worthwhile benchmarks instead bs 3d whatever or super pi?

Quote:

---

Originally Posted by bateslabel

Is it too hard to post worthwhile benchmarks instead bs 3d whatever or super pi?

---

Is it too hard to post something constructive or positive? All the tests they have done are very universal and I like to be able to compare my computer to Conroe. Furthermore, I appreciate anything they can give us (no other site has provided information like this about these chips), and I'm sure more will come.

what would be a "worthwhile benchmark"?

Here ya go:

http://www.xtremesystems.org/forums/...ad.php?t=98091

That is as real an application as it gets. I don't think anyone else can get 165FPS average in Half Life 2: Lost Coast timedemo unless they also own a Conroe chip.

Here's my shot at giving an apples to apples comparison. It would seem some of the A64 benches posted aren't running tight timmings.
http://members.cox.net/gtabor1/2.7.JPG

I dont get it.
I got this results:
8135 pts
Test 1= 3.7
Test 2= 8.1

Running a opteron 165 at 3006mhz (mem at 250).
How come you beat my 3ghz opteron in game 1 and I beat you in test 2 when your cpu was at 2.6? Is memory bandwidth that important for 3dmark cpu test? What am I doing wrong in here?

if u dont know then dont mind but the cpu test on 05 is bugt ass hell .
we @ belgim tested , get evry time some sort of other score , some time it can change with even a 1000punts ;) .
so plz dont use the 05 cpu tests as good compare

Quote:

---

Originally Posted by freecableguy

3.6Ghz vs. 2.66Ghz isn't fair either....there is no fair.....let it go.

---

Very nice work guys very nice indeed!