Like I said, I ran Prime95 for 24hrs without errors so I consider it pretty stable. And yes I consider 60C a bit high but I'm using retail heatsink/fan/pad....not even any thermal grease. I think it's mainly because I have the voltage so high. My memory is running at stock speed so I didn't need to increase the ddr voltage. Does anyone know how to even increase ddr voltage on an asus p4s333?
How do you increase the Voltage for RDRAM? (TH7II-RAID MB) I have heard that RDRAM is very picky about voltage. Has anybody heard differ?
-x86
-x86
1.6A Northwood @ 2.26 GHZ on a Asus P4S333 with King Max DDR 2700 Ram with a core voltage of 1.6 (stock HSF) has been running prime 95 for 10 1/2 hours now without any issues. Prime 95 would give me errors after about an hour with the voltage at 1.575 but seems fine at 1.6. CPU temp has hovered between 49C and 52C for the duration of the test.
Strangely though if I set the FSB any higher at all (no matter what CPU voltage I use) the system will not even post. Maybe its a memory issue.
later
Strangely though if I set the FSB any higher at all (no matter what CPU voltage I use) the system will not even post. Maybe its a memory issue.
later
After 48 hours of prime 95 my MSI845Ultra ARU with the p4_1.6a running at 1.58ghz my temps are reading 60c. That is very high, but msi has there tems frigged up i think. If not its still stable as hell and im going to get another hs/f soon to see if i can restrain the heat a lil bit.
My 1.6a using the stock Intel Fan with an Epox 4SDA+ @1.575 volts
Will post at 2.4 Ghz, 150 Mhz bus, 1.65 volts, but won't boot windows.
@ 2.34 Ghz, 146 Mhz bus, I can run Quake 3, Sandra, etc. with no problems, but it fails Prime 95 after a few minutes.
@ 2.24 Ghz 140 Mhz bus I can run Prime 95 & compress Mpeg2 files all day, no errors.
Also that puts my Samsung DDR memory @ 175 Mhz, Cas 2 (2-3-3) which gives me around 2600/2500 on the Sandra memory bandwidth test.
IMHO if it doesn't run Prime 95 overnight (with the covers on), then it's not a stable overclock.
Will post at 2.4 Ghz, 150 Mhz bus, 1.65 volts, but won't boot windows.
@ 2.34 Ghz, 146 Mhz bus, I can run Quake 3, Sandra, etc. with no problems, but it fails Prime 95 after a few minutes.
@ 2.24 Ghz 140 Mhz bus I can run Prime 95 & compress Mpeg2 files all day, no errors.
Also that puts my Samsung DDR memory @ 175 Mhz, Cas 2 (2-3-3) which gives me around 2600/2500 on the Sandra memory bandwidth test.
IMHO if it doesn't run Prime 95 overnight (with the covers on), then it's not a stable overclock.
As read above, I couldn't get my computer to run prime without errors at very high clockspeeds. I also latter raun into errors on the 2.07Ghz OC. I finally figured out the problem: Memory. I had TOO MUCH of it (1GB). After taking out half, I got total stability on 2.2GHz. These are the POSSIBLE causes:
1. Having all 4 banks full of memory will not be able to provide the votage needed to each stick because the whole bank is alotted a certain maximum amount. When OCing, as we all know, memory needs more voltage, but once the maxium has been reached, the memory will die out, and errors will occur.
2. The heat caused by having all 4 congested banks of memory filled is tremendous, considering the fact that this is RDRAM, which is a major heat producer. (But to counter this theory, I have a 80mm fan right next to the memory banks, and my memory temps while having all 4 filled was 39C).
3. The more RIMMs are used, the less likely they will run at higher clock speeds. In other words, the probability of 4 RIMMS running at a higher clock speed is much less than the probability of only two RIMMS running at a higher clock speed.
I took a lot of time writing this so please give me commentary.
Thanks
-x86
1. Having all 4 banks full of memory will not be able to provide the votage needed to each stick because the whole bank is alotted a certain maximum amount. When OCing, as we all know, memory needs more voltage, but once the maxium has been reached, the memory will die out, and errors will occur.
2. The heat caused by having all 4 congested banks of memory filled is tremendous, considering the fact that this is RDRAM, which is a major heat producer. (But to counter this theory, I have a 80mm fan right next to the memory banks, and my memory temps while having all 4 filled was 39C).
3. The more RIMMs are used, the less likely they will run at higher clock speeds. In other words, the probability of 4 RIMMS running at a higher clock speed is much less than the probability of only two RIMMS running at a higher clock speed.
I took a lot of time writing this so please give me commentary.
Thanks
-x86
I'm glad I could help. It's interesting that extra RAM caused your problems...
BTW, did you find out if you can increase the voltage on RDRAM? If you can up the voltage a bit you might beable use all the RAM that you have.
It's interesting that extra RAM caused your problems... BTW, did you find out if you can increase the voltage on RDRAM? If you can up the voltage a bit you might beable use all the RAM that you have.
Very interesting and thanks for posting the cure for your problem. I guess I won't be running out and grabbing another 512MB of RDRAM. I am surprised it took you so long to try this though. That would have been the first place I would have started. Also, did you try running all the RDRAM (1GB) at 3X instead of 4X? You may be able to keep it all in at 3X and overclock to a 133 fsb and have the RDRAM at spec 400. Of course you give up bandwidth going this route.
Did you ever figure out how to increase the RDRAM voltage on the TH7II? I didn't think it was possible.
Did you ever figure out how to increase the RDRAM voltage on the TH7II? I didn't think it was possible.
Thanks for the reply. I don't even know why I bought 1GB of memory. How the heck can you POSSIBLY USE IT ALL?! Now I have a good reason for getting rid of it . LarryJoe, why do you need another 512MB? How do you use it up?
I was thinking that it was the RAM all along, as you can see in my previous posts about running 4 RIMMs as opposed to 2. But I really regret that it took me so long.
If I decreased the bandwidth of the RAM what is the purpose for OCing then? Can anybody provide a reason as to why you would underclock the RAM when you are OVERclocking? It essentially defeats the purpose. To answer your question, yes, I did try underclocking it, but that didn't work. I still haven't found a way to increase the voltage of the RAM, but I am pretty sure that the MB provides all four banks a CONSTANT amount of Volts it can provide. To give an example:
The voltage of RDRAM is 2.5V, so all four banks are capable of supplying 2.5 X 4 = 10V to the whole bank. If you decrease the number of banks being used, then the two banks will have the headroom required in reaching higher frequencies since only 5V of the 10 is utilized. This is ONE of the MANY theories I have.
-x86
. LarryJoe, why do you need another 512MB? How do you use it up? I was thinking that it was the RAM all along, as you can see in my previous posts about running 4 RIMMs as opposed to 2. But I really regret that it took me so long.
If I decreased the bandwidth of the RAM what is the purpose for OCing then? Can anybody provide a reason as to why you would underclock the RAM when you are OVERclocking? It essentially defeats the purpose. To answer your question, yes, I did try underclocking it, but that didn't work. I still haven't found a way to increase the voltage of the RAM, but I am pretty sure that the MB provides all four banks a CONSTANT amount of Volts it can provide. To give an example:
The voltage of RDRAM is 2.5V, so all four banks are capable of supplying 2.5 X 4 = 10V to the whole bank. If you decrease the number of banks being used, then the two banks will have the headroom required in reaching higher frequencies since only 5V of the 10 is utilized. This is ONE of the MANY theories I have.
-x86
I wasn't exactly saying underclock the RDRAM, just set it to 3X at 133 and you will be running it in spec. This is not underclocked. But you are right, you would lose memory bandwidth. Some people would rather see a higher cpu clock speed and sacrafice a little memory bandwidth.
I tried running at 2.67/400 vs. 2.5/500 (my ram can't do 533) and the results were very similar in the benches like 3DMark. Sandra's CPU benches were obviously better at 2.67 with the memory scoring better at 500.
To answer your questions about need for 1gb of memory and overclocking in general -
It is all about bragging rights my friend. Do we REALLY need this much speed? Do we even notice the speed difference between 1.8 and 2.4?? Nope. Did I really need to upgrade from a PIII@1.333 to a P4@2.67? Nope, as I can't really tell the difference all that much. Sure it is fast, but not like going from a Voodoo1 to a Vodoo2 SLI!
I tried running at 2.67/400 vs. 2.5/500 (my ram can't do 533) and the results were very similar in the benches like 3DMark. Sandra's CPU benches were obviously better at 2.67 with the memory scoring better at 500.
To answer your questions about need for 1gb of memory and overclocking in general -
It is all about bragging rights my friend. Do we REALLY need this much speed? Do we even notice the speed difference between 1.8 and 2.4?? Nope. Did I really need to upgrade from a PIII@1.333 to a P4@2.67? Nope, as I can't really tell the difference all that much. Sure it is fast, but not like going from a Voodoo1 to a Vodoo2 SLI!
RalfHutter
Diamond Member
- Dec 29, 2000
- 3,202
- 0
- 76
As I've posted above, my box is running a 1.8A @ 2.3 Ghz (128 SB), with 2 x 512Mb ( that's 1 Gb) of Corsair PC2700 DDR RAM @ 340Mhz, CAS2, 2.5v. My OC is 100% stable at these setting. I have no trouble running 1Gb of DDR RAM (of course it's the good Corsair stuff - which uses Samsung chips, BTW). Maybe it's an RDRAM/1Gb issue?
I have no idea why people spend 100's of dollars on ram they will never use. 1Gb? Dizzam...I was always buying 512Mb when I built a pc, I just built this P4 box a few months ago and went with 256Mb DDR, and I still don't use it all. I always have ~100Mb unused. I don't think even massive use of PSP would use a Gig of ram.
"It is all about bragging rights my friend.: ----must be right
"It is all about bragging rights my friend.: ----must be right
<< As I've posted above, my box is running a 1.8A @ 2.3 Ghz (128 SB), with 2 x 512Mb ( that's 1 Gb) of Corsair PC2700 DDR RAM @ 340Mhz, CAS2, 2.5v. My OC is 100% stable at these setting. I have no trouble running 1Gb of DDR RAM (of course it's the good Corsair stuff - which uses Samsung chips, BTW). Maybe it's an RDRAM/1Gb issue? >>
Not to flame you, but the 1GB of RDRAM was PRECISELY the point that I was trying to make on my post above. My theories circulate around that facts of RDRAM.
-x86
RalfHutter
Diamond Member
- Dec 29, 2000
- 3,202
- 0
- 76
Nate420 - I do Photoshop and video on this box and I've seen memory usage around 600megs more than a few times. I can actually use this RAM and it's my money to spend so I did. I have no regrets.
<< << As I've posted above, my box is running a 1.8A @ 2.3 Ghz (128 SB), with 2 x 512Mb ( that's 1 Gb) of Corsair PC2700 DDR RAM @ 340Mhz, CAS2, 2.5v. My OC is 100% stable at these setting. I have no trouble running 1Gb of DDR RAM (of course it's the good Corsair stuff - which uses Samsung chips, BTW). Maybe it's an RDRAM/ 1Gb issue? >>
Not to flame you, but the 1GB of RDRAM was PRECISELY the point that I was trying to make on my post above. My theories circulate around that facts of RDRAM.
-x86 >>
I know you're running RDRAM, that's why I specifically mentioned that my RAM is DDR. I just wanted to point out the fact that it's probably an RDAM issue, not a "1 gig in general" issue.
<< << As I've posted above, my box is running a 1.8A @ 2.3 Ghz (128 SB), with 2 x 512Mb ( that's 1 Gb) of Corsair PC2700 DDR RAM @ 340Mhz, CAS2, 2.5v. My OC is 100% stable at these setting. I have no trouble running 1Gb of DDR RAM (of course it's the good Corsair stuff - which uses Samsung chips, BTW). Maybe it's an RDRAM/ 1Gb issue? >>
Not to flame you, but the 1GB of RDRAM was PRECISELY the point that I was trying to make on my post above. My theories circulate around that facts of RDRAM.
-x86 >>
I know you're running RDRAM, that's why I specifically mentioned that my RAM is DDR. I just wanted to point out the fact that it's probably an RDAM issue, not a "1 gig in general" issue.
P4B266, NW 1.6a, 2x256 PC2700 Samsung DDR, P4 Volcano HSF.
Running at 2.13 Ghz, default vcore
Ran stable for ~2 weeks under normal usage (RTCW, MOH:AA).
40C temp idling, 48C max temp under load
Prime95 ran for about 4 hours before I stopped it.
Running at 2.4 Ghz, vcore set in bios to 1.700 (asus probe reports actual vcore ~1.65)
idles at 40C, 56C max temp under load
Prime95 failed after 31 min.
Do I need better cooling, or am I just pushing this poor lil' northwood too far?
I'm going to put it back to 2.13 Ghz at default vcore and see what Prime95 does overnight.
Azathoth
----------------------------------------------------------------------------
Disclaimer: I work for Intel. The views expressed above are my personal opinions and in no way represent those of Intel.
Running at 2.13 Ghz, default vcore
Ran stable for ~2 weeks under normal usage (RTCW, MOH:AA).
40C temp idling, 48C max temp under load
Prime95 ran for about 4 hours before I stopped it.
Running at 2.4 Ghz, vcore set in bios to 1.700 (asus probe reports actual vcore ~1.65)
idles at 40C, 56C max temp under load
Prime95 failed after 31 min.
Do I need better cooling, or am I just pushing this poor lil' northwood too far?
I'm going to put it back to 2.13 Ghz at default vcore and see what Prime95 does overnight.
Azathoth
----------------------------------------------------------------------------
Disclaimer: I work for Intel. The views expressed above are my personal opinions and in no way represent those of Intel.
Your temperatures seem alright. I personally think that it is the memory. Have you taken memory temps?
-x86
-x86
bump
i'm pretty much in the same boat as azatouth, but its too late to post my stats now. will do it 2mrow...
i'm pretty much in the same boat as azatouth, but its too late to post my stats now. will do it 2mrow...
Epox 4BDA2+, 1.8A - vcore at 1.63v, pci and agp locked at spec, running at 2.4GHz with Intel HS/F. Current temperature is 38 degrees C. Gets as high as 53 C during extreme stress testing.
I am not OC'ing my video card but I was getting some oddities in video and couldn't get the one game I occasionally play to get completely loaded without a spontaneous reboot. Was also having a disappearing keyboard (MS Internet Kybd Pro) after restarts. Keyboard has both USB and PS2 connectors with both plugged in. Put USB keyboard support in bios to disabled (default is Auto) and solved that problem.
Now here comes the strange one (at least to me), I upped the AGP voltage by .10 volts. Wallah, system is silky smooth, cpu temperature dropped to it's current level ( was 41 at same conditions), temperature swings are much more gradual (upward that is), all programs including my game work great! If you are getting odd reboots, try it.
I'll have to run Prime for awhile but I know she'll be fine. I made it through quite a number of loops in Prime95 before finding my AGP cure.
I am not OC'ing my video card but I was getting some oddities in video and couldn't get the one game I occasionally play to get completely loaded without a spontaneous reboot. Was also having a disappearing keyboard (MS Internet Kybd Pro) after restarts. Keyboard has both USB and PS2 connectors with both plugged in. Put USB keyboard support in bios to disabled (default is Auto) and solved that problem.
Now here comes the strange one (at least to me), I upped the AGP voltage by .10 volts. Wallah, system is silky smooth, cpu temperature dropped to it's current level ( was 41 at same conditions), temperature swings are much more gradual (upward that is), all programs including my game work great! If you are getting odd reboots, try it.
I'll have to run Prime for awhile but I know she'll be fine. I made it through quite a number of loops in Prime95 before finding my AGP cure.
i'm no expert in p4 o/c since i don't actually have one yet, but maybe you should mess with your memory timings azathoth, maybe it's the ram thats holding you back.
I'm sure glad I read this thread. This is my first overclocking. I'm beginning to think my under overclocking can be improved by changing the setting for my memory. I have the MSI 845Ultra, 1.6A, Crucial 512Mb (1 stick) with core voltage @1.65v running rock solid @1.87 Ghz. I'd prefer not to do the wire mod to get higher core voltage until the last resort. I got it up to 1.89Ghz running Sandra, but couldn't reboot. Increasing the core voltage would more than likely let me increase the FSB up from 117 (up from set/ fallback 100 and 1.6Ghz). Appreciate any advice you care to offer. I don't have a clue about the procedure, settings, etc. concerning the memory. Thanks, Pat.
I forgot to mention I 'm using the Sunflower HSF in place of the Intel. My CPU temp 59-62F and Chassis temp 89 to 92F
can be improved by changing the setting for my memory. I have the MSI 845Ultra, 1.6A, Crucial 512Mb (1 stick) with core voltage @1.65v running rock solid @1.87 Ghz. I'd prefer not to do the wire mod to get higher core voltage until the last resort. I got it up to 1.89Ghz running Sandra, but couldn't reboot. Increasing the core voltage would more than likely let me increase the FSB up from 117 (up from set/ fallback 100 and 1.6Ghz). Appreciate any advice you care to offer. I don't have a clue about the procedure, settings, etc. concerning the memory. Thanks, Pat. I forgot to mention I 'm using the Sunflower HSF in place of the Intel. My CPU temp 59-62F and Chassis temp 89 to 92F
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 23K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter