If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below. |
|
|
Thread Tools | Display Modes |
#1
|
|||
|
|||
Problems overclocking Aus P4C800
"Maveric" skrev i en meddelelse ... Hi group, I have run into some trouble overcloking my system. All components are new and are : Asus P4C800 motherboard Corsair XMS CMX512A-3200C2 (two sticks) DDR400 Cas2.0 memory Pentium IV 2,4B (133x18) Processor Q-Tec Gold 550 PSU the usual suspects (2x HD, 1 x DVD-ROM, 1 x ZIP, 1x floppy & 5 casefans) The Asus board from the word go sets the memory speed to DDR333 (166Mhz). I started from there and increase the core clock upwards. All went fine (POST + load of Windows 98SE and 1 run of 3D Mark 2001 SE) until i reached 156Mhz FSB (2808Mhz Core clock). No POST possible above this speed until i lowered the memory to DDR 266 (133Mhz). I continued from here until i hit another brick wall at 166Mhz FSB (2988Mhz Core clock). No POST beyond that ? As i understand it (and have experienced it) overclocking stops first under hgh load. If you push further then the OS may fail to load, if you still continue THEN POST may fail ........ right ? That's not what is see. (i increased the FSB in small steps of 1 Mhz at a time) Also why does the FIXED memory speed react on the increasing FSB speed, the northbridge isolates the two.... not ? The memory should be capable of DDR400 (200Mhz) but at high FSB speeds it had even trouble to stay at DDR266 (133Mhz) what gives ? I feel the CPU has more potential but or the motherboard is holding me back or the memory is flakey/incompatible ? Anyone ? Greetings, Frans Baetsen Have you tried increasing the voltage to the CPU and the memory (Vdimm) ? I'm running the same CPU as you, at 165MHz FSB, and have had to increase the Vcore voltage a little, to make it stable at this speed. My memory is Corsair PC 3500 (433 MHz) running at 413 MHz. Below their specs, but with agressive memory timings, which required an increase of the voltage. Trying to run at default voltage resulted in a complete lockup where I couldn't even get into the BIOS to change the settings. Had to do a jumper reset. Dunno if any MBs allow you to set the memory frequency completely independent of the FSB. On my Chaintech 9EJL2 I have two alternatives, sort of a low and high setting, and both vary along with the FSB. At 165 MHz FSB, I get the choice between 330 and 413 MHz memory frequency. My CPU seems unhappy to go further in speed, at least without increasing the Vcore voltage to levels I'm uncomfortable using, so I have more or less settled for its current speed of just about 3 GHz, which is a decent increase of about 25 %. Regards, Erling Groes-Petersen |
#2
|
|||
|
|||
The memory speed is NOT fixed. It's a ratio of processor speed. Although the
BIOS may tell you that you're setting a memory speed, what you're really doing is setting a ratio of memory speed to FSB speed. I think the P4C800 mobo gives you three choices of memory dividers (400) which is really 1:1, (333) which is really 4:5, (266)which is really 2:3. If the memory can't hang in at 1:1 with 156 FSB, I'd be suspicious of it. In BIOS, disable Memory Acceleration Mode, which is trying to make use of a feature your processor does not have. Try one stick at a time and see what happens. Try relaxing the timing. If you still can't get memory to run faster, think about sending it back. At 166 FSB (which is a 25% overclock), you may be hitting your processor limit, but the memory should easily run at the (400) 1:1 setting and tight timing. Hope this helps. "Erling Groes-Petersen" wrote in message k... "Maveric" skrev i en meddelelse ... Hi group, I have run into some trouble overcloking my system. All components are new and are : Asus P4C800 motherboard Corsair XMS CMX512A-3200C2 (two sticks) DDR400 Cas2.0 memory Pentium IV 2,4B (133x18) Processor Q-Tec Gold 550 PSU the usual suspects (2x HD, 1 x DVD-ROM, 1 x ZIP, 1x floppy & 5 casefans) The Asus board from the word go sets the memory speed to DDR333 (166Mhz). I started from there and increase the core clock upwards. All went fine (POST + load of Windows 98SE and 1 run of 3D Mark 2001 SE) until i reached 156Mhz FSB (2808Mhz Core clock). No POST possible above this speed until i lowered the memory to DDR 266 (133Mhz). I continued from here until i hit another brick wall at 166Mhz FSB (2988Mhz Core clock). No POST beyond that ? As i understand it (and have experienced it) overclocking stops first under hgh load. If you push further then the OS may fail to load, if you still continue THEN POST may fail ........ right ? That's not what is see. (i increased the FSB in small steps of 1 Mhz at a time) Also why does the FIXED memory speed react on the increasing FSB speed, the northbridge isolates the two.... not ? The memory should be capable of DDR400 (200Mhz) but at high FSB speeds it had even trouble to stay at DDR266 (133Mhz) what gives ? I feel the CPU has more potential but or the motherboard is holding me back or the memory is flakey/incompatible ? Anyone ? Greetings, Frans Baetsen Have you tried increasing the voltage to the CPU and the memory (Vdimm) ? I'm running the same CPU as you, at 165MHz FSB, and have had to increase the Vcore voltage a little, to make it stable at this speed. My memory is Corsair PC 3500 (433 MHz) running at 413 MHz. Below their specs, but with agressive memory timings, which required an increase of the voltage. Trying to run at default voltage resulted in a complete lockup where I couldn't even get into the BIOS to change the settings. Had to do a jumper reset. Dunno if any MBs allow you to set the memory frequency completely independent of the FSB. On my Chaintech 9EJL2 I have two alternatives, sort of a low and high setting, and both vary along with the FSB. At 165 MHz FSB, I get the choice between 330 and 413 MHz memory frequency. My CPU seems unhappy to go further in speed, at least without increasing the Vcore voltage to levels I'm uncomfortable using, so I have more or less settled for its current speed of just about 3 GHz, which is a decent increase of about 25 %. Regards, Erling Groes-Petersen |
#3
|
|||
|
|||
On Sat, 28 Jun 2003 15:05:36 +0200, "Erling Groes-Petersen"
wrote: Hello Erling, Have you tried increasing the voltage to the CPU and the memory (Vdimm) ? Yes i have, for 166Mhz i need 1,700V for 158Mhz 1,625V is alright. For the memory i need 2,65 . If i go higher stability for 'Memtest-86' tests 1-6 increases but test 7 goed totaly wrong. I'm running the same CPU as you, at 165MHz FSB, and have had to increase the Vcore voltage a little, to make it stable at this speed. My memory is Corsair PC 3500 (433 MHz) running at 413 MHz. Lucky beep You can get your memory speed above 333, the asus board wil not let you do that, only with 800Mhz C type CPU's You shure have not the Asus P4C800 Below their specs, but with agressive memory timings, which required an increase of the voltage. Trying to run at default voltage resulted in a complete lockup where I couldn't even get into the BIOS to change the settings. Had to do a jumper reset. Hmmmm same here. Strange, a motherboard should be able to at east get into the BIOS without memory ...... not ? Dunno if any MBs allow you to set the memory frequency completely independent of the FSB. On my Chaintech 9EJL2 I have two alternatives, sort of a low and high setting, and both vary along with the FSB. At 165 MHz FSB, I get the choice between 330 and 413 MHz memory frequency. Sounds better then i can choose My CPU seems unhappy to go further in speed, at least without increasing the Vcore voltage to levels I'm uncomfortable using, so I have more or less settled for its current speed of just about 3 GHz, which is a decent increase of about 25 %. That's what i was aiming for also. Some further testing showed the memory is the problem. My type of Corsair memory does not work well with my Asus board it seems. It was not certified by Asus but since I was not planning running it a max. spec (DDR400) I figured that it would not matter.... now it does Regards, Erling Groes-Petersen Thanks for the reply Greetings, Frans |
#4
|
|||
|
|||
On Sat, 28 Jun 2003 21:58:54 -0400, "Frank Weston"
wrote: Hello, The memory speed is NOT fixed. It's a ratio of processor speed. Although the BIOS may tell you that you're setting a memory speed, what you're really doing is setting a ratio of memory speed to FSB speed. I think the P4C800 mobo gives you three choices of memory dividers (400) which is really 1:1, (333) which is really 4:5, (266)which is really 2:3. That would explain a lot. There is nothing in the BIOS or manual to suggest this. If the memory can't hang in at 1:1 with 156 FSB, I'd be suspicious of it. In BIOS, More testing on my part showed that my type of Corsair memory is not compatible with my Asus board. It was not certified by Asus but since i was not planning on running it at max spec (DDR400) i thought i would be OK .... not thus disable Memory Acceleration Mode, which is trying to make use of a feature your processor does not have. Still off, i was planning to play with that when i found a combination of FSB and memory that worked stable. Try one stick at a time and see what happens. Tried that, one of the sticks goes 2 MHZ FSB (160Mhz) further when running single channel. The other is the bad one (152Mhz max). But both are not good. If i apply your theory then this memory craps out at well below 133Mhz actual memory speed ...... bummer Try relaxing the timing. If you still can't get memory to run faster, think about sending it back. At 166 FSB (which is a 25% overclock), you may be hitting your processor limit, but the memory should easily run at the (400) 1:1 setting and tight timing. I have been playing with that. 158Mhz FSB with CL2.0 3 4 is the best i can get there (2.0 2 3) is not stable and at 166 Mhz CL2.0 2 3 does run Memtest-86 OK. When testing memory bandwidth the CL value does nothing, CL2.0 is as fast as CL3.0, but the RAS & RAS2CAS values do give a performance boost. Hope this helps. Yes it does, time to get better memory Greetings, Frans |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Gigabyte GA-NSNXP-939 Problems | Paul | AMD x86-64 Processors | 5 | December 31st 04 06:59 AM |
Sapphire Radeon 9800SE + Shuttle machine = Problems | Jiri Kuukasjärvi | General | 2 | September 12th 04 11:17 AM |
Looking for some medium-to-extreme overclocking advice | Cyde Weys | General | 4 | July 2nd 04 04:03 AM |
(CROSSPOST): Stability problems | Frodo | General | 2 | December 21st 03 12:59 AM |
Major Computer Problems | Toronto Garage Door Company | General | 20 | November 13th 03 10:41 PM |