San Diego and Venice comparison at same clock speed

**EMC2** · 05-10-2005, 05:13 PM

Q's....

1) Did you disable Thermal Throttling in the BIOS?

2) Do you have interleave enabled?

3) Do you have DynIdleCntr disabled?

4) Are you running 1T?

5) Have you played with RdWrQueueBypass, ByPassMax, & IdleCycleLimit?

6) Have you disabled DynIdleCycleCnt?

7) Do you have anything set to "Auto" in the BIOS?

Peace

**ben805** · 05-10-2005, 05:17 PM

Originally Posted by agenda2005

Can you run both of them at 10X multi and the same frequency, HTT and memory timming? Because lowering multiplier has been know to cause loss of efficiency, especially in Sandra.
The only way we can clear this up is when we have both of them at thier highest multiplier posible, hence 3500+ is required for comparison with 3700+.
Secondly, can you change your fan RPM so that your 3200+ run with low rpm and hence higher temps, with 3700+ running with high fan speed/rpm and hence lower temps. Try and balance the temps as much as you can to help clarify the temperature/heat issues.

hmm....yes I can change RPM, but my brother's dang Venice was only about 1~2'C off with the fan spinning at 1800~2000rpm, I don't think it's a good idea to run it any lower speed just incase I roast his Venice, well maybe I should

In the mean time, there's another San Diego owner had the same heat problem as mine and he said after removing the IHS it went from 56'C to 42'C!!!! check this out He also gained 100mhz!!! holy smoke...

**Gsus_Freak** · 05-10-2005, 05:21 PM

Well when you incress cashe size, you also incress latency. I bet if you ran a cashe test you would see that is the case here. However, in gamming tests, massive cashe swaping isn't going to be as much as a problem. Run a farcry becnch an it should swing in your favor.

We had the same delima with the clawhammer and newcasle cores with the atrifitial bechmarks. That is why the guys at [H]ard OCP never use syinthetic benches.

**saaya** · 05-10-2005, 05:27 PM

interesting... thx a lot for those results!

**EMC2** · 05-10-2005, 05:30 PM

--- Gsus ---

Look at the Sandra Cache/Mem tests he posted

Yes, artificial benchies can be misleading...

**perkam** · 05-10-2005, 05:35 PM

Very simple explanation for all of this...run both at the same speed (the SD with divider) with UTT or BH-5 and you'll see the 3700+ SD outshine the Venice...

Perkam

**ben805** · 05-10-2005, 05:37 PM

Originally Posted by dmo580

This is a bad benchmark. Raise your hand those of you who have seen the 3700+ vs the 3800+ where the 3700+ whoops @$$ in virtually all the benchmarks? Ok, enough said. Do the test on the same system plz. Kthx.

Bad benchmark? totally agreed...these numbers are NOT what I expected, and NOT what I wanted to see either, I wish my SD can whoop as much azz as that SD that I saw

<raising hands...here here here> and that's the main reason why I bought the SD in the first place, if I didn't see that SD whoopazz comparison then I wouldn't of spent the extra 144 bucks....dammit!! not that I'm blaming that comparison but this is just not what I expected!! and YES they are being tested on the same system!

Originally Posted by EMC2

Q's....

1) Did you disable Thermal Throttling in the BIOS?

2) Do you have interleave enabled?

3) Do you have DynIdleCntr disabled?

4) Are you running 1T?

5) Have you played with RdWrQueueBypass, ByPassMax, & IdleCycleLimit?

6) Have you disabled DynIdleCycleCnt?

7) Do you have anything set to "Auto" in the BIOS?

Peace

1- yes
2- interleave always disable
3- dynamic cycle always disable
4- always 1T, my 512LA is superpi 32m stable up to 337.5 @2.5-4-3-7 1T and can bench at 345 @2.5-4-3-7 1T.
5- those are always being set at 8x, 4x, and 16clks.
6- dynamic counter always disable
7- yes, only the drive strength, but with bios 414-3 the auto always default to weak...i believe it's 7..

All settings are the same on both processor, except when I attempted to push for the highest superpi 32M and even then...I only change FSB, Multiplier and Vcore, I haven't touch the memory settings ever since I got it stable at 337.5 @2.5-4-3-7 1T.

**agenda2005** · 05-10-2005, 05:37 PM

ben,
Can you download and run sciencemark 2 (molecular dynamic and primodia). Don't use default setting, but use large number of silicon atoms that will take several minutes. This is a memory bandwith hungry application and will surely expose the deficiency of any CPU architecture.

**ben805** · 05-10-2005, 05:42 PM

Originally Posted by perkam

Very simple explanation for all of this...run both at the same speed (the SD with divider) with UTT or BH-5 and you'll see the 3700+ SD outshine the Venice...

Perkam

YOU READ MY MIND!! This morning I had already ordered a pair of Mushkin Redline XP4000 and had it 2nd day air to me, if my SD isn't defective and heat didn't impact the performance that much, then I really hope the SD would perform slightly better on UTT/BH?

**Gsus_Freak** · 05-10-2005, 05:45 PM

man, where are guys digging up those winbound chips? Would love to get my greedy little hands on them. Anyway, I will look forward to the results. This thread is making my desision to get the SD or the X2 even harder

. AMDzone.net pushed there X2 to 2.7 on air!

Ok ill stop hijacking the thread now. Anyway, have you tried a gamming bench yet to see if there is any real world difference? Would like to see that. Anyway, enjoy your new ram...muchkin is the best stuff man...I have some bh-6's now, solid like a rock.

**ben805** · 05-10-2005, 05:55 PM

Originally Posted by agenda2005

ben,
Can you download and run sciencemark 2 (molecular dynamic and primodia). Don't use default setting, but use large number of silicon atoms that will take several minutes. This is a memory bandwith hungry application and will surely expose the deficiency of any CPU architecture.

Ok, I saw two version...32 and 64bit, it said 2.0 Final but which one should I use? I opened the 32bit one...and it has a bunch of Blas/Cipher/Primordia/Mol/Stream benches.....never run this before so which one do you want me to run? I'm clueless here...

And is there anything else that I can download and run so maybe and hopefully I can beat my brother's Venice?

Originally Posted by Gsus_Freak

man, where are guys digging up those winbound chips? Would love to get my greedy little hands on them. Anyway, I will look forward to the results. This thread is making my desision to get the SD or the X2 even harder

. AMDzone.net pushed there X2 to 2.7 on air!

Ok ill stop hijacking the thread now. Anyway, have you tried a gamming bench yet to see if there is any real world difference? Would like to see that. Anyway, enjoy your new ram...muchkin is the best stuff man...I have some bh-6's now, solid like a rock.

thanks....have never play with BH? stuff before so hopefully the redline would get along with my SD.

**Gsus_Freak** · 05-10-2005, 06:16 PM

BH love to be OC'ed. You can crank some seious voltage through them for great high frequecy cas timmings.

**agenda2005** · 05-10-2005, 06:17 PM

Originally Posted by ben805

Ok, I saw two version...32 and 64bit, it said 2.0 Final but which one should I use? I opened the 32bit one...and it has a bunch of Blas/Cipher/Primordia/Mol/Stream benches.....never run this before so which one do you want me to run? I'm clueless here...

And is there anything else that I can download and run so maybe and hopefully I can beat my brother's Venice?

If you are on normal Windows XP, you will need to run the 32-bit version. Run the Mol Dyn and Primordia using large number of atoms, say 1024 atoms.
1024 atoms will take all your 1MB L2 cache and therefore shows the advantage of more cache.

**viccyran** · 05-10-2005, 06:44 PM

Ben, the problem for your temps might involve bad contact with the IHS. The only sol'n I know of is to operate on your CPU

**ben805** · 05-10-2005, 06:55 PM

Originally Posted by agenda2005

Can you run both of them at 10X multi and the same frequency, HTT and memory timming? Because lowering multiplier has been know to cause loss of efficiency, especially in Sandra.
The only way we can clear this up is when we have both of them at thier highest multiplier posible, hence 3500+ is required for comparison with 3700+.
Secondly, can you change your fan RPM so that your 3200+ run with low rpm and hence higher temps, with 3700+ running with high fan speed/rpm and hence lower temps. Try and balance the temps as much as you can to help clarify the temperature/heat issues.

Hey Agenda, I read your post again and it got me thinking... IF my brother were to bought a 3500 instead of the 3200 Venice, THEN he came back and do another comparison with my SD running at the same multiplier on same FSB, should I expect his Venice to get waxed this time around?!? Or should my San Diego getting waxed even worst?? IF...just IF...my SD have the table turned this time and end up outperforming his 3500 Venice WITH the same multiplier and same FSB, doesn't it make the 3500 Venice a suckier cpu than the 3200 Venice? Since I'm still using TCCD at the moment, the only way I can get the most performance out of these memory and cpu is to drop the multiplier and ramp up the FSB to over 300, and run the ram on 1T to get the best out of them, so in this case....I would still end up dropping the multiplier on the 3500 back down to 9x. Running on air, they both top out at around 2850~2862, which is pretty close....strange...

Originally Posted by viccyran

Ben, the problem for your temps might involve bad contact with the IHS. The only sol'n I know of is to operate on your CPU

I can't, phase change unit is on the way soon...I prefer to leave the IHS on otherwise I'll risk crushing the core

Beside, it's only fair to take off the Venice too if I were to strip off mine, I think my brother is going to kill me if I cut his baby wide opened....

Maybe if somebody happen to get their hands on both Venice and San diego, I'm curious to see what the numbers look like if they are willing to put them on similar comparison test, if the result comes out as we all hope and expected, then I know for sure I've got a defective SD and I'll be RMA this crap back....

**agenda2005** · 05-10-2005, 07:03 PM

Originally Posted by ben805

Hey Agenda, I read your post again and it got me thinking... IF my brother were to bought a 3500 instead of the 3200 Venice, THEN he came back and do another comparison with my SD running at the same multiplier on same FSB, should I expect his Venice to get waxed this time around?!? Or should my San Diego getting waxed even worst?? IF...just IF...my SD have the table turned this time and end up outperforming his 3500 Venice WITH the same multiplier and same FSB, doesn't it make the 3500 Venice a suckier cpu than the 3200 Venice? Since I'm still using TCCD at the moment, the only way I can get the most performance out of these memory and cpu is to drop the multiplier and ramp up the FSB to over 300, and run the ram on 1T to get the best out of them, so in this case....I would still end up dropping the multiplier on the 3500 back down to 9x. Running on air, they both top out at around 2850~2862, which is pretty close....strange...

Well I can say exactly what will happen. This is strange.

However, a little correction with the sciencemark. Run Primordia with default setting and molecular dynamics with 1024 atoms. These applications love caches alot.

**ben805** · 05-10-2005, 07:18 PM

just tried the Sciencemark thing...instant BSOD again!! LOL Prime95 failed instantly as well...I think I was a bit optimistic wishing I could prime this thing at 318 x9 on air

**ImportantAwareness89** · 05-10-2005, 07:32 PM

Ok, in the following article we have a 3800+ Venice (2.4 GHz) vs a 3700+ SD (2.2 GHz).

This thread was about same clock speed, and it showed Venice winning. However, if a 3800+ goes against a 3700+, the Venice should have an EVEN GREATER advantage right?

http://www.compressorcooling.com/htm...go_vs__ve.html

Now imagine both at the same clock speed. I say the SD kicks the Venice around EVEN MORE. In 3D Mark 2003, the Venice wins, but I say 2003 is not as good as a bench as 2001. 2001 has more CPU reliance, which is why I feel its better.

AND NO, you did not test your stuff on the same system. You tested on 2 systems with "similar components." Tell me that's good enough when you do medical testing. The rhesus monkey doesn't die from drug testing and so we can skip human trials all along? Right. You do realize a simple DEFRAG makes things a lot more different. When I picked up my old system, it was doing 2900 in 3D Mark 2k1, but with a defrag it shot up roughly 150 points.

**trakslacker** · 05-10-2005, 08:07 PM

Originally Posted by dmo580

AND NO, you did not test your stuff on the same system. You tested on 2 systems with "similar components." Tell me that's good enough when you do medical testing. The rhesus monkey doesn't die from drug testing and so we can skip human trials all along? Right. You do realize a simple DEFRAG makes things a lot more different. When I picked up my old system, it was doing 2900 in 3D Mark 2k1, but with a defrag it shot up roughly 150 points.

actually from the way he is talking, I believe he did in fact test both CPUs on the EXACT same system. Switched the CPUs out between the same physical components in the same case.

**ben805** · 05-10-2005, 08:17 PM

Originally Posted by dmo580

Ok, in the following article we have a 3800+ Venice (2.4 GHz) vs a 3700+ SD (2.2 GHz).

This thread was about same clock speed, and it showed Venice winning. However, if a 3800+ goes against a 3700+, the Venice should have an EVEN GREATER advantage right?

http://www.compressorcooling.com/htm...go_vs__ve.html

Now imagine both at the same clock speed. I say the SD kicks the Venice around EVEN MORE. In 3D Mark 2003, the Venice wins, but I say 2003 is not as good as a bench as 2001. 2001 has more CPU reliance, which is why I feel its better.

AND NO, you did not test your stuff on the same system. You tested on 2 systems with "similar components." Tell me that's good enough when you do medical testing. The rhesus monkey doesn't die from drug testing and so we can skip human trials all along? Right. You do realize a simple DEFRAG makes things a lot more different. When I picked up my old system, it was doing 2900 in 3D Mark 2k1, but with a defrag it shot up roughly 150 points.

what made you think that these two cpu were not being tested on an identical system??

but instead you keep saying they were running on a different....well...one of your imaginary and 'different system' that I've been hiding behind the desk covered with...perhaps a magic carpet?

maybe you're right...I should of run some defrag or maybe chkdsk...or something....how can I not think of that?!?! sheesh...my bad

and God forbid I wish the AMD angel would show up tonight and sprinkle some magic dust on my processor so I wouldn't look like an azz infront of my brother that I'd spent $344 on a processor only to got outperformed by his Venice, oh yea...MR Santa Claus....don't you dare take away the spare pc that was hiding behind my desk...pleaaase....

I'm really sorry and apologize that these benchmarks disappointed you, or anyone else, as much as it does to me....but it's just a comparison between two processors that me and my bro got, and they do not represent any other Venice or San Diego out there either, and who know...perhaps I'm just unlucky and got a defective one, or maybe I fell on my head and forgot that I'd actually ran these two cpu on different setup, and yea...the San Diego rules..and it SHOULD kick the crap out of the Venice!!

satisfied?

**ben805** · 05-10-2005, 08:45 PM

Just to clear some things up, as any experience overclockers here should know, at some point...overclocking your hardware WILL corrupt your Windows due to instable memory timing and such, especially pushing your memory at much higher speed and tighter timing than it's rated for, and if you're running a RAID-0 like mine....you know what kind of chances you got on hosting the Raid array. I have several acronis and ghost images made just for this reason....quick recovery and to get everything up and running in less than 20min. And that is why I did not have the same desktop icons and wallpaper when I benched the Venice and San Diego, but since someone is accusing me of running these two cpu on different machines, I will go the extra miles and put this sort of unnecessory suspicion to rest, I'm going to load up the ghost image that is identical to the one that I ran on Venice and run some of these benches again.

Mr dmo580, I hope you'd be prepare for another wave of shock_and_awe, yes sir...

**Geforce4ti4200** · 05-10-2005, 08:49 PM

This makes no sense to me but if its true and word gets out, no ones gonna buy a Diego then

Good thing I got a 3000+ Venice thats cheaper and faster than Diego but it cant be right

**charlie** · 05-10-2005, 08:58 PM

If anyone want to get rid of their "dog" diego's, I'll give you a 3200 Venice and a pat on the back, lol....

Any takers?

C

**Vincentvega18** · 05-10-2005, 09:02 PM

Could this be the bandwidth bug weve been seeing with tccd and the san diego...... maybe between the 2 theres no such thing as 1t with current bios. Just a thought.

**zakelwe** · 05-10-2005, 09:16 PM

Originally Posted by Vincentvega18

Could this be the bandwidth bug weve been seeing with tccd and the san diego...... maybe between the 2 theres no such thing as 1t with current bios. Just a thought.

I would imagine it is something along those lines, I guess possibles are

a) For some reason a bad SD chip in this one case ( unlikely but impossible )
b) Some BIOS / memory / motherboard / setting combionation ( likely )
c) All SD's perform worse than Venice ( extremely unlikely ).

Thanks for posting the results, will be interesting to see if this is a one off.

Regards

Andy

Thread: San Diego and Venice comparison at same clock speed

Thread Tools

Search Thread

Rate This Thread

Display

Bookmarks

Bookmarks

Posting Permissions