- May 10, 2009
- 226
- 0
- 0
Hello everyone. I would like to ask for your opinion on a simple matter, but I have to tell you the whole story first.
I just got a new i5 2500K system and it's running flawlessly overclocked to 4.5GHz @ 1.3V (I got fairly lucky with these voltages from what I've read), along with an extreme-budget motherboard (MSI P67A-C43 [B3]), which I'm liking as well so far. Here is the trick:
I had 4 gigs (2x2) of leftover OCZ Gold 1600MHz triple-channel 1.65V RAM from the previous system, so I decided to run with it until I get something more permanent. However, the RAM was only stable at a maximum frequency of 1333MHz (though I was able to lower the timings to 7-7-7-20 from the stock 8-8-8-24), and if put at 1600MHz it would freeze randomly, most of the time while watching videos or surfing the web. What is also interesting is that even after the BIOS update both CPU Z and the BIOS itself would only recognize the RAM to be rated @ 1066 max frequency, even though I am absolutely sure it is meant to run at 1600. Knowing the apparent "weaknesses" of Sandy Bridge, in the sense of it not officially supporting RAM above 1333MHz, out of curiosity I put some old 2GB stick of Geil RAM rated at 1333MHz and you know what? Not only did the board recognize it properly, I could also overclock it to 1600MHz @ 1.65V and get it to run perfectly stable! At this point, I had already ordered 8GB (2x4) HyperX 1600MHz 1.65V from Kingston. When these new sticks arrived, something interesting happened...
It turns out that the geniuses I ordered the RAM from sent me one wrong stick, which was, instead of the original HyperX I ordered, HyperX Blu, the budget version of the one I wanted. But that wasn't my biggest problem at this point. The biggest issue was that this Blu version was rated at 1.7-1.9V, an extremelly high voltage I would never dare to run on a Sandy Bridge platform. So I put the right stick into the mobo, tried it at 1600MHz and 1.65V, worked like a charm. The board recognized it as 1333, which was also fine as this is the max frequency Sandy Bridge officially supports. I tried to lower the timings to squeeze a bit more performance out of it, but it wouldn't allow any adjustments at all. So, out of my damn curiosity again, I figured that the board, which in the BIOS doesn't recommend more than 1.65 volts would be smart enough not to destroy itself if paired with incompatible RAM, so I put the Blu stick in the board as well to run on dual channel. I was right. It did not fry. Also, according to the BIOS, the two sticks were completely the same, both apparently rated at only 1.5V, but on the first start they were both automatically set at 1.65. They worked well at 1600MHz, but that wasn't enough for me, lol...
I set the frequency to 1866MHz, left the volts on auto. Something extremelly weird happened: the system wouldn't boot, it would just start the fans and violently shut itself off after 3 seconds, and it would probably have continued to do this in many loops had I not turned the power off. I figured no problem, since I had already seen this happen when I set incorrect timings. What freaked me out is that I felt IMMENSE heat coming from the PC, so I thought the motherboard set the volts too high and that I was doomed... Luck smiled upon me again, and after a CMOS reset everything worked fine.
What I finally want to ask any person who was kind enough to read through all of this is: could the board have set the voltage too high, even exceeding its own safety limits? Maybe the heat feeling was just an illusion, maybe I overreacted. From someone experienced, I would like to hear if you would recommend trying to overclock these sticks any further with tighter timings, and if you believe I should leave the voltage on auto, going against my instincts.
Thank you so much to whomever has been so great as to read this pile of text I typed here to show you the greater picture. I will appreciate every opinion or statement. Thank you again.
I just got a new i5 2500K system and it's running flawlessly overclocked to 4.5GHz @ 1.3V (I got fairly lucky with these voltages from what I've read), along with an extreme-budget motherboard (MSI P67A-C43 [B3]), which I'm liking as well so far. Here is the trick:
I had 4 gigs (2x2) of leftover OCZ Gold 1600MHz triple-channel 1.65V RAM from the previous system, so I decided to run with it until I get something more permanent. However, the RAM was only stable at a maximum frequency of 1333MHz (though I was able to lower the timings to 7-7-7-20 from the stock 8-8-8-24), and if put at 1600MHz it would freeze randomly, most of the time while watching videos or surfing the web. What is also interesting is that even after the BIOS update both CPU Z and the BIOS itself would only recognize the RAM to be rated @ 1066 max frequency, even though I am absolutely sure it is meant to run at 1600. Knowing the apparent "weaknesses" of Sandy Bridge, in the sense of it not officially supporting RAM above 1333MHz, out of curiosity I put some old 2GB stick of Geil RAM rated at 1333MHz and you know what? Not only did the board recognize it properly, I could also overclock it to 1600MHz @ 1.65V and get it to run perfectly stable! At this point, I had already ordered 8GB (2x4) HyperX 1600MHz 1.65V from Kingston. When these new sticks arrived, something interesting happened...
It turns out that the geniuses I ordered the RAM from sent me one wrong stick, which was, instead of the original HyperX I ordered, HyperX Blu, the budget version of the one I wanted. But that wasn't my biggest problem at this point. The biggest issue was that this Blu version was rated at 1.7-1.9V, an extremelly high voltage I would never dare to run on a Sandy Bridge platform. So I put the right stick into the mobo, tried it at 1600MHz and 1.65V, worked like a charm. The board recognized it as 1333, which was also fine as this is the max frequency Sandy Bridge officially supports. I tried to lower the timings to squeeze a bit more performance out of it, but it wouldn't allow any adjustments at all. So, out of my damn curiosity again, I figured that the board, which in the BIOS doesn't recommend more than 1.65 volts would be smart enough not to destroy itself if paired with incompatible RAM, so I put the Blu stick in the board as well to run on dual channel. I was right. It did not fry. Also, according to the BIOS, the two sticks were completely the same, both apparently rated at only 1.5V, but on the first start they were both automatically set at 1.65. They worked well at 1600MHz, but that wasn't enough for me, lol...
I set the frequency to 1866MHz, left the volts on auto. Something extremelly weird happened: the system wouldn't boot, it would just start the fans and violently shut itself off after 3 seconds, and it would probably have continued to do this in many loops had I not turned the power off. I figured no problem, since I had already seen this happen when I set incorrect timings. What freaked me out is that I felt IMMENSE heat coming from the PC, so I thought the motherboard set the volts too high and that I was doomed... Luck smiled upon me again, and after a CMOS reset everything worked fine.
What I finally want to ask any person who was kind enough to read through all of this is: could the board have set the voltage too high, even exceeding its own safety limits? Maybe the heat feeling was just an illusion, maybe I overreacted. From someone experienced, I would like to hear if you would recommend trying to overclock these sticks any further with tighter timings, and if you believe I should leave the voltage on auto, going against my instincts.
Thank you so much to whomever has been so great as to read this pile of text I typed here to show you the greater picture. I will appreciate every opinion or statement. Thank you again.
Last edited: