Benchmark

[Guest]

If you're looking for a new case check out the Antec 1200. I just moved into this case, a VERY nice case/.

 

What are you running for memory? PC6400 Pc8000?

 

3 fans???

 

I have

 

(3) 120mm Front fans(intake)

(2) 120mm rear fans(exhaust)

(1) 200mm top (exhaust)

(1) CPU fan - Zalman 9500

(1) 60mm Northbridge fan

(2) 120mmm Internal fans

 

each GTS8800 extreme has its own exhaust fan as well

[YoDoug®]

quote:

---

(3) 120mm Front fans(intake)

(2) 120mm rear fans(exhaust)

(1) 200mm top (exhaust)

(1) CPU fan - Zalman 9500

(1) 60mm Northbridge fan

(2) 120mmm Internal fans

---

Did you have to bolt the puter down so it doesn't take off

[FP1]

JP, you have enough fans in that thing to make hurricane Katrina seem like just another windy day!

 

This is my home system. Did the whole rig for $680 from Newegg(No OS)back in April. Nothing too fancy/expensive. It has some potential for upgrading.

 

Cooler Master Centurion 5 case

(1) 80mm front fan

(1) 120mm rear fan

Antec Earthwatts 500w PS

Gigabyte GA-P35-DS3L MB

Intel C2D E6750 2.66 GHz

(2) 1Gb sticks of Kingston KHX6400D2LLK2

(1)MSI 8600GT video card

(1)WD 320G/16M/SATA2

 

I'll probably just do a good cpu cooler and overclock for now. I have always wanted to jump into a quadrro but the fact is this little 8600GT for $85 has been working out fine for my current Mastercam needs in the home/office.

[Guest]

quote:

---

Did you have to bolt the puter down so it doesn't take off [big Grin]

---

Nah, the TOP fan keeps it down on the table

[Guest]

Running that CPU, you might be better off and get better results using memory that is timed to 667 instead of the 800 mhz you've got..

 

333 Mhz x 4 = 1333 FSB

 

333 Mhz x 8 = 2.6 Ghz

 

your memory timings are being down clocked to meet your CPU

 

667 x 2 = 1333

 

So your memory timing are 2:1

 

If you go with PC5400 memory you could maintain a 1:1 RAM:FSB ratio.

 

However, if you choose to try to bump your FSB to 400, keep your multiplier at 8x's and you can then run your memory at he full 800 mHz speed.

 

That would give you an effective 3.2 Ghz.

[FP1]

quote:

---

However, if you choose to try to bump your FSB to 400, keep your multiplier at 8x's and you can then run your memory at he full 800 mHz speed.

 

That would give you an effective 3.2 Ghz.

---

This is what I am going to try for.

[FP1]

Just tried it at 400 X 8 = 3.2 Ghz and 800 MHz for the memory. Benchmarked in 5 min 35 secs at a temp of 50/63C. What about my Northbridge temp? I don't see a way to monitor that from my BIOS.

[Guest]

Try downloading Speedfan.

 

It will depend if your motherboard has a sensor for it

http://www.softpedia.com/progDownload/Spee...nload-4103.html

[bogusmill]

Nice time John! Where can I download the online stopwatch? I've often wanted to use that instead of a hand held stop watch.

 

So you have a CPU with less cache than mine and your running at 3.2 vs. my 3.8 yet your 9 seconds faster than me. What could be the difference making you faster? OS + 2Gb RAM maybe?

 

I loaded Vista Ultimate SP1 x64 this weekend but I didn't try the MC benchmark yet. However I was able to get Catia x64 to load and I was able to rotate a part smoothly with all performance settings maxed (including AA) that won't rotate smoothly in XP x86 until I reduce settings.

 

I'm looking at an 8Gb RAM purchase soon and I learned about something else that might be nice boost. Not saying what but you'll hear about it if it works.

 

MLS has asked if I could help him build a new PC and another friend wants to build the sucessor to my board, the Rampage Extreme with lots of RAM so things are getting faster all the time. The more we learn the better for all of us!

[Guest]

www.online-stopwatch.com

 

C'mon you wont share

 

How to completely explain not sure, extra RAM, 10k drives. What is the Mhz on your RAM, while at the 3.2 OC speed I can run my RAm at 1:1 @ 800Mhz.

[bogusmill]

My RAM is running at 1129Mhz or so. I'll put up a shot of my CPU-Z tonight but 1:1 is as slow as you can go. If you look at your memory (dividers) they are actually multipliers. Mine is something like 6/5.

 

Don't be surprised if I retake the lead tonight. As we've seen before superior CPU speed and RAM amount & speed didn't get me faster than a CPU with more L2 cache. I have Vista loded on a non RAIDed Raptor but that won't help the benchmark anyway.

 

quote:

---

C'mon you wont share

---

I will if it helps. If it doesn't lots of time will be wasted by some forum members researching this.

[bogusmill]

Here is my CPU-Z and Everest screen captures. As you can see in both I'm running RAM at 1140Mhz and it's at a 6/5 multiplier.

 

 

 

The Vista 64 bit OS helped by 5 seconds. Maybe 8 gigs of RAM will improve my time more than 4 seconds John. I'll let you know but it will be a few weeks.

 

[FP1]

Put in my new heat sink and fan last night. It is an Arctic Freezer Pro 7. It reduced my temps by 4C on idle and 10C while loaded. The directions for installation aren't so good so I am posting more realistic directions for our members:

 

1. remove old heat sink & fan.

2. Clean thermal paste off of CPU.

3. apply fresh thermal paste to CPU

4. Insert heat sink so the four plastic lock pins are in the mounting holes on the mother board.

5. press down on locking pins in opposite corner until they "click".

6. One of them for sure won't click and you will either break or bend the little plastic expander.

7. Remove heat sink, fix or replace plastic pin.

8. Repeat steps 2 through 5. Now your pins should be locked.

9. Apply band-aid to cut finger.

10. Now take a closer look at the heat sink. You were so diligent with making sure you didn't F-up the pins you have mounted the heat sink in the wrong orientation and the fan won't fit.

11. Remove the heat sink and repeat steps 2 through 5 again

12. Apply another band-aid, this time to a different finger.

13.Mount the fan while shaking your head in a side to side fashion in dis-belief over the previous steps.

14. Re-assemble the rest of the computer

15. Clean up blood trails and your good to go!

[Guest]

Doing some stress testing this morning on a 3 Ghz OC

 

 

Ran a bench this am, I'll be doing some more as well

 

[Guest]

Looks like unless I want to really start doing some "exotic" stuff, I am about maxed out @ 3:56 min/sec

 

I added my new 4 gigs of RAM in did away with the paging file and clocked in @ 4:04

 

So I arrived at an overclock to 3.2 Ghz, that tested all fours cores for almost 2 hours straight with all 4 running at 100% my highest Temp on the cores remained at 60°C, I idle at 30-33°C

 

I am going to stress test overnight to verify the stability of the overclock but it appears quite stable at the moment, I figure at worse at this point I might have to give it another .1 volts.

 

For the curious, yes I did bump up to 400Mhz X 9 for a 3.6 Ghz, I ran right at 70°C at 2 minutes of stress testing the cores, so on an air cool only I abandoned those settings.

[Guest]

bogusmill,

 

Have you run Prime95 on your system to stress test it?

 

You are correct at your speed you "should" be a bit fast er than I am and times are show a difference.

 

I stress tested with Prime95 overnight last night for 13 hours the cores and memory ran at 100% stress.

 

 

I did not get a single calculation error in the test for the duration. So from my initial testing, which had my CPU voltage at 1.4312, I now have it set at 1.35 volts and I am completely stable and not running hot. SO I am wondering if you are not perhaps experiencing some calculations errors along that way and that is forcing recalculations.

 

If you are running a quad-core remember to copy prime95 to 4 different folders and to run 4 different instances, each set to run on it's own core. You might find you are not completely stable.

 

Just at thought

[Guest]

I can not as yet stay stable at 3.6 Ghz but I did run Benchmark, 3:32

 

 

My system stays stable to do everyday things but as soon as I engage the stress testing program, BSOD

I am at 1.45v on my core, using the 15% theory on upping the voltage and my VID of 1.3, I could max out at 1.49v

 

I think what all of this proves is anyone needing SERIOUS number crunching and using ANYTHING less than a 1333FSB chip and 12 megs of L2 cache is going to be in for LONG waits.

[MotorCityMinion]

"I think what all of this proves is anyone needing SERIOUS number crunching and using ANYTHING less than a 1333FSB chip and 12 megs of L2 cache is going to be in for LONG waits."

 

Heh, or get another Cam package.

[iowajim62]

quote:

---

Heh, or get another Cam package.

---

[bogusmill]

John, yes I've run Prime 95 but not for as long as you did. Can you run Everest? My >10000 score in write is rare from what I've seen. For CPU cooler I used the Thermalright Ultra 120 Extreme. It's one of best air coolers made and it goes together with screws instead of those push pins we all hate so much.

 

The thing I was looking at that may speed things up is CUDA. It's a process that uses your GPU for stream calculations and it doesn't have to be video. I don't know if a toolpath is that type of processing but the program (MCX3) would have to include code to enable that so no help from Nvidia's GPU.

 

BTW what were you doing up at 12:22AM Sat. night (Sun. morning) John?

[Guest]

That was actually in the afternoon.

 

12:22 am is waay past my bedtime

 

I'll see if I can run Everest and put the scores tonight.

 

This Q6600 really IS quite a overclocker. A LOT of potential. It really is amazing the different the 1600 FSB makes in program performance

[bogusmill]

quote:

---

12:22 am is waay past my bedtime

---

AM or PM. I was just guessing, now we know.

 

You have up to 9x available, I only have up to 8x. 475fsb X 9 wow, if only it would work! I couldn't find what MB your using.

[Guest]

for the 3.2Ghz Setting, I am using 400Mhz x 8 times

 

On the 3.6Ghz settings I tried 400 Mhz x 9X, I can get it to boot at 1.45volts but I can not get more than a minute or 2 into the Prime95 testing. I might try creeping up on another .25 volts or so, when I get few minutes to see if I can get it stable.

 

I have stayed away from anything above 400 Mhz to allow my RAM to run at 1:1 which "should" be the best for it.

[bogusmill]

You didn't say whats your MB.

 

quote:

---

I have stayed away from anything above 400 Mhz to allow my RAM to run at 1:1 which "should" be the best for it.

---

Actually thats slowing you down. Here is a review of my MB but on the 4th page it gives a detailed description of straps and memory multipliers and why 1:1 is as slow as you can go.

 

http://www.anandtech.com/mb/showdoc.aspx?i=3208

[Guest]

quote:

---

You didn't say whats your MB.

---

I am running as Asus P5N-D