Multiplier & 12V line "slipping" - ***? - Overclock.net

de_baser · 4 Weeks Ago

Hey!

I've been running a nice little setup with Q9450 and gigabyte Ga-x48t-dq6 for 18 months, overclocking happily and stabily. Now i found it was time to get new RAM, and invested in the new HyperX T1 (the ones with the huuuge heatsinks) DIMMs. Turns out they weren't very happy to work with my mobo and i eventually had to RMA to get the older generation 1800mhz 4GB hyperX set.

Plugged these puppies in yesterday, and everything seems fine. These even perform faster on everest tests than the glitchy T1's (which were rated at 2000mhz). However, i've come across a weird problem with these that i have not seen earlier on my setup: Once every five minutes, CPU-Z & OCCT suddenly read the multiplier at 6x, before jumping back up to the default, static 8x after a few seconds (this doesn't read in everest, as everest doesn't read the multi in real-time). At first i considered this a weird software bug, but it turns out this is also readable in the power supply! The 12V line drops at the same time to 5-6V and glithces around before restoring itself.

This behaviour is unsettling and seems to hinder my search for stability at nice, high speeds. I've never reduced the multiplier earlier, but it turns out the T1's had a tendency to do so, as they came with XMP profiles by default. Meaning that when the board reset on a failed clock, XMP forced the board to reboot to 500x6 in order to run at 2000 MhZ. The new, old-gen HyperX's don't have XMP or anything, but it is with these that the drop to 6x occasionally happens. My old corsair 1333MhZ sticks don't cause this problem - i'm not 100% sure if this problem is isolated to RAM alone.

Does anybody have any idea what is going on? I am quite confused. Considering to flash the BIOS to see if that will work, but i don't know if it's worth the bother. Is this an issue of another pair of incompatible sticks? Any other suggestions?

Clockadile Dundee · 4 Weeks Ago

Bizzaro. Fill out your system specs in the User CP.

twisted_58 · 4 Weeks Ago

try running memtest 1 stick at a time then 1 slot at a time with a good stick. that will tell you if its the ram or the mobo. be sure to do all of this at stock settings

de_baser · 4 Weeks Ago

A small overhaul, flash and clean of the whole rig is a classic way of attempting to fix things - like blowing on the NES cartridge. And believe it or not, it actually fixed the problem with the jumpy multiplier. Reads well in OCCT and CPU-Z now.

However, my 12V line & CPU don't seem to like eachother in OCCT. I run whatever stress tests there, and every couple seconds the load level (100%CPU load) drops to some random number, often to zero. Reading the diagrams, i see that the 12V scintillates with this as well. In one of these random drops i occasionally crash out of the test, even though the machine seems to be running happily.

I know the 12V reading from software is never accurate, but i don't remember OCCT to be this jumpy. Is this normal?

4 Weeks Ago	#1 (permalink)
de_baser New to Overclock.net Join Date: Apr 2008 Posts: 6 Rep: 0 Unique Rep: 0 Trader Rating: 0	Multiplier & 12V line "slipping" - ?* Hey! I've been running a nice little setup with Q9450 and gigabyte Ga-x48t-dq6 for 18 months, overclocking happily and stabily. Now i found it was time to get new RAM, and invested in the new HyperX T1 (the ones with the huuuge heatsinks) DIMMs. Turns out they weren't very happy to work with my mobo and i eventually had to RMA to get the older generation 1800mhz 4GB hyperX set. Plugged these puppies in yesterday, and everything seems fine. These even perform faster on everest tests than the glitchy T1's (which were rated at 2000mhz). However, i've come across a weird problem with these that i have not seen earlier on my setup: Once every five minutes, CPU-Z & OCCT suddenly read the multiplier at 6x, before jumping back up to the default, static 8x after a few seconds (this doesn't read in everest, as everest doesn't read the multi in real-time). At first i considered this a weird software bug, but it turns out this is also readable in the power supply! The 12V line drops at the same time to 5-6V and glithces around before restoring itself. This behaviour is unsettling and seems to hinder my search for stability at nice, high speeds. I've never reduced the multiplier earlier, but it turns out the T1's had a tendency to do so, as they came with XMP profiles by default. Meaning that when the board reset on a failed clock, XMP forced the board to reboot to 500x6 in order to run at 2000 MhZ. The new, old-gen HyperX's don't have XMP or anything, but it is with these that the drop to 6x occasionally happens. My old corsair 1333MhZ sticks don't cause this problem - i'm not 100% sure if this problem is isolated to RAM alone. Does anybody have any idea what is going on? I am quite confused. Considering to flash the BIOS to see if that will work, but i don't know if it's worth the bother. Is this an issue of another pair of incompatible sticks? Any other suggestions?

		Overclock.net - Overclocking.net > Intel > Intel CPUs
Multiplier & 12V line "slipping" - ?*

4 Weeks Ago	#4 (permalink)
de_baser New to Overclock.net Join Date: Apr 2008 Posts: 6 Rep: 0 Unique Rep: 0 Trader Rating: 0	A small overhaul, flash and clean of the whole rig is a classic way of attempting to fix things - like blowing on the NES cartridge. And believe it or not, it actually fixed the problem with the jumpy multiplier. Reads well in OCCT and CPU-Z now. However, my 12V line & CPU don't seem to like eachother in OCCT. I run whatever stress tests there, and every couple seconds the load level (100%CPU load) drops to some random number, often to zero. Reading the diagrams, i see that the 12V scintillates with this as well. In one of these random drops i occasionally crash out of the test, even though the machine seems to be running happily. I know the 12V reading from software is never accurate, but i don't remember OCCT to be this jumpy. Is this normal?

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)