Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
All We Know About Maximizing CPU Related Performance
- Thread starter Tutor
- Start date
- Sort by reaction score
Ok, I'm doing the test at 150/100/13 with Vcore @ 1.325 and so far in my 20 min OCCN Test, it's still up and running. I will try 160/100/13. Then see what happens... Thanks...
EDIT: I had to up the Vcore to 1.35V on all of them and then I ran my test. It didn't shut down this time it just gave me a quacking duck (that scared the crap out of me while I was working on my other system - LOL !!!), but when I saw that there was an ERROR message on OCCN, I gathered that the quacking noise meant there was some kind error. I will up it two more notches here shortly and do another test. BTW, I put (in bold) what I have setup in my total values so far in my BIOS to help you see what I have. Hope that helps, thanks...
Last edited:
It's so low because your CPU multiplier is that low. Remember, you are going to do Memory in Phase 2.
----------
Ok, I'm doing the test at 150/100/13 with Vcore @ 1.325 and so far in my 20 min OCCN Test, it's still up and running. I will try 160/100/13. Then see what happens... Thanks...
I predict you will need more VCore....
You are correct. I currently have the Vcore set to 1.36250V and it kicked it for 43 min into the OCCT test and then QUACKED on me. So now I'm up to 1.37500V. We'll see if this will work. If not then I'll notch it by 2 more and go from there. I'm trying to stay as low on the Vcore as I can go as I remember Tutor saying that not even he goes beyond 1.35 as he knows the risk of frying the CPUs is there if you go beyond that point.
1) If 1.40 in my Vcore works, would that be what I would use in my permanent Vcore settings if thats the only Voltage that does work, and is that safe?
2) Prior to this test I had the Vcore set at 1.35625V and it BSOD on me. Here's the pic. I was able to get it in the nick of time as I would have missed it had I not turned my head at the right time. It happened so fast, but I got it. What does this screen tell you?
Attachments
Near bottom, there are several numbers in Hex....The first one is the key, 0x0A indicates unstable memory controller, suggested fix for this is either to increase VTT or VCore if that doesn't work.
You'll see lots of people pushing VCore higher, you may not need to go quite so high because you are underclocking, and that is one of the touted benefits. I suspect that to get his very highest overclocks, Tutor had to turn up Vcore beyond 1.35, but only he can answer that directly. High VCores theoretically shorten life of the chip, but it depends on how good your cooling is as well. Yours are under water, so you should be running substantially cooler. I know for a fact that my overclocked chips run 10 to 15 C Cooler than my chips in my real Mac Pro run. Even 1.45 is very commonly deployed. At a notch below 1.45 my chips idle at 29-30 Celsius (ambient is 26) and IntelBurnTest at 58-72 Celsius (depending if in socket 1 or 2). For everyday use running my simulations, max temps are around 55-60 Celsius.
Last edited:
Thanks for that bit of education on the BSOD stuff. Man, you really have to be sitting in front of the screen to take that pic, because if you miss that info for more than 5 to 10 seconds, you're SOL and the next thing you know, you're system reboots. Now I'm sitting next to my system (using another one), but my back is to it, because I'm working and don't have time to sit and watch everything. I was just fortunate enough to catch it as I was turning around to check the screen and my camera was within reach. Is there a way to get that to info another way if I miss the BSOD when it happens (since there's a 10 sec window of opportunity)?
Anyway, I know I'm digressing a bit, but viewing your BIOS (in your Hardware Health Configuration menu), what are your CPU 0 and CPU 1 Temps? I ask this because I wanted to find out the Celsius temp differences between the two CPUs (as I know that one runs a little hotter than the other). On idle (15 min after start up, with my fans set @ 1,500 rpm) mine looks like this:
CPU 0 Temp: 27C
CPU 0 PWM Temp: 38C
System 0 Temp: 33C
CPU 1 Temp: 28C
CPU 1 PWM Temp: 37C
System1 Temp: 35C
If you had a moment to check it, I was just curious. Of course when I first start up my system (first thing in the morning) it's 4C cooler than what I just gave you. That's why I give it 15+ min to warm up so the temps (on idle) are more accurate. I've heard of others posting 10C differences between CPU 0 & CPU 1 on their SR-2 systems. That's why I'm asking. I haven't had that issue. The most I've seen was a 1C to 4C difference, but typically it's 1C to 2C higher on CPU 1.
Within my OCCT Torture Tests the highest temps that I've seen my system go up to is 65C (mind you, I have my whole system water cooled (CPUs, GPU and Mobo). I haven't seen it go any higher. So I think I'm okay in that area when it comes to temps. Also, I have two Rads (one 480 and one 360 and they're both fully fanned up, push/pulling). Lastly, during my tests I have my Corsair Excalibur's cranked up all the way to 2000 rpm (typically though I'm at around 1,500 to 1,600 rpm).
Okay, now back to upping the VCore (or VTT). I will start with the VCore as I haven't gone up to 1.40V yet. I'm 2 notches away from that happening and will start that test later today as I tend to business matters today. I will let you know how that goes later on this afternoon. BTW, thanks for the PM...
Last edited:
You can turn off the auto reboot after system failure. That might help although mine occasionally shutsdown anyway. Blue screen viewer is name of program to see bsod.
Where would I find this Off/On switch? Thanks...
PS - about the temp thing, I know you're doing a simulation test, but after you're done (when you have a moment) I would be interested in finding out. I'll explain why later...
Where would I find this Off/On switch? Thanks...
PS - about the temp thing, I know you're doing a simulation test, but after you're done (when you have a moment) I would be interested in finding out. I'll explain why later...
Lol I'm not familiar enough with windows to walk you thrpugh it, but ita under advanced options in the system configuration control panel.
My CPU sockets are 5 degrees or so apart. I have my pumps and fans turned down so you can't hear them...fans at 600 rpm. That's why temps go to 69 or 70 on occn linpack. If I turn everything up it's I'm the low 60s, just too loud for my tastes. Temps same range as yours at idle, both CPUs about the same, pwms about the same. Ambient temps are 26 c in day
Thanks for laughing at me...
(no worries, not insulted I'll get over my tears... ).I say this because someone here made a comment about me using too much tubing because of the way I set up my looped system:
http://www.xtremesystems.org/forums...-SR-2-Mac-Hak-Pro-Build-YES-IT-FITS-!!!/page2
(Read post #34). It seems that since you're CPUs are 5C in temp difference each other, then my point is now more verified that separating the loop and not feeding one (water) loop directly into the other (i.e., CPU0 to CPU1), but rather setting it up the way I did helps a out a bit as I replied to RCB_Bex (in Post #35) in the same link I just provided (above). Here's the orchestration of my water looped system:
Swiftech Maelstrom Dual Bay/MCP35X2 (housing w/dual pumps in series)>4 Rad>CPU0 (1st CPU)>Mobo Waterblock>3 Rad>CPU1 (2nd CPU)>GPU Waterblock>back to Dual Bay Res & round & round she goes...
So with that setup, I'm now able to see a significant change in the CPU temps being very close to each other as they have their own dedicated Rads feeding into each CPU separately. I've even seen my CPUs (from time to time) at the exact same temps (i.e.; CPU0 = 28C & CPU1 = 28C on idle). You may want to consider that with your next SR-2 build. Just a thought as my findings are starting to hold true; even with my Mobo & GPU water blocks that are also part of that looped setup...
I'll hit you back up with my other OCCT Torture Test (TT) later this afternoon and let you know how it's going. Thanks again for all the input...
Last edited:
Sigh.. I have to set Vcore to 1.36V in order to run stably with a blck of 160. This seems too high..
Also, has anyone had any experience with their second GPU "disappearing"? While I'm playing a game, the screen goes black. If I hit the reset button to reboot, only 1 GPU shows up. I have to totally power down to get the other GPU to show up again.
Also, has anyone had any experience with their second GPU "disappearing"? While I'm playing a game, the screen goes black. If I hit the reset button to reboot, only 1 GPU shows up. I have to totally power down to get the other GPU to show up again.
Last edited:
No. 1.36 is not too high for a 160 bclk, although not sure which chip you have.
For your gpu problem you may need to turn up the IOH
----------
Thanks for laughing at me...
I say this because someone here made a comment about me using too much tubing because of the way I set up my looped system:
http://www.xtremesystems.org/forums...-SR-2-Mac-Hak-Pro-Build-YES-IT-FITS-!!!/page2
(Read post #34). It seems that since you're CPUs are 5C in temp difference each other, then my point is now more verified that separating the loop and not feeding one (water) loop directly into the other (i.e., CPU0 to CPU1), but rather setting it up the way I did helps a out a bit as I replied to RCB_Bex (in Post #35) in the same link I just provided (above). Here's the orchestration of my water looped system:
Swiftech Maelstrom Dual Bay/MCP35X2 (housing w/dual pumps in series)>4 Rad>CPU0 (1st CPU)>Mobo Waterblock>3 Rad>CPU1 (2nd CPU)>GPU Waterblock>back to Dual Bay Res & round & round she goes...
So with that setup, I'm now able to see a significant change in the CPU temps being very close to each other as they have their own dedicated Rads feeding into each CPU separately. I've even seen my CPUs (from time to time) at the exact same temps (i.e.; CPU0 = 28C & CPU1 = 28C on idle). You may want to consider that with your next SR-2 build. Just a thought as my findings are starting to hold true; even with my Mobo & GPU water blocks that are also part of that looped setup...
I'll hit you back up with my other OCCT Torture Test (TT) later this afternoon and let you know how it's going. Thanks again for all the input...
Lots of mythology in water cooling. Loop order really doesn't matter, and I doubt it's why you see no delta in yours. For instance, in my loop I run to CPU 1 first (the hot one), then CPU 0, yet there is still a delta. I've run it the other order as well, same difference in delta t. Again, lots of info about how loop order doesnt really matter significantly with adequate flow rates
I have IOH set to 1.4V. I was planning on lowering this actually but haven't had time to test it yet. This is with X5680s. I'm going to try to change the PCIe freq from 103 to 100 to see if that helps with the GPU disconnect.
CPU's vary tremendously when you begin to push them above the factory speed setting. Got two SR2's - one with a set of CPUs that requires about 6 to 8 more steps of CPU Vcore than the other set requires to yield the same performance level. Neither requires as much as yours do however. To bad we can't "test drive" them before we buy them.
Yeah, I feel better about it looking at the Vcores used in this thread.
Maybe I can find some unlocked X56XXs around here somewhere...
Yeah, I was about to say, not everyone can run them as low as on the Vcore as Tutor's. All we can do is try...
Your suspicion is correct. For the 40,100 and 40,051 scores (freq = 2483; 2483/13 ~ 191) at 191 bclk I pumped somewhere between 1.425 and 1.45 Vcore in them.
But when I used overclocking, higher Vcore yielded me a score of only 37,837 and that bench garbled my drive and the test data - Note that my system's speed shows as only 26 MHz in my geekbench profile, and as I recalled, I had to redo my HD.
Wait, (and again, please forgive my ignorance), but now I am confused?? I thought you've been advocating not to go above 1.35 V while all the time you have been able to keep that steady GB score (of 40,100 with a Vcore that is lower than 1.35 V), so is that still the case, or are you only reaching those high GB scores when you're using those higher voltages? If that's the case then, what is your current GB score with your current voltages? Thx...
Last edited:
Well, it's just my findings. Also, when spoke with five other SR-2 owners that have their water looped setup directly linked together (one CPU to the next CPU) with an 8C to 10C CPU temp difference, that just added to my own reasoning, as I have the 480 Rad directly feeding cool coolant first into CPU 0 and then mobo water block, and as it comes out warmer it then goes into the other 360 Rad and comes out cool again and directly feeds into CPU 1 then the GPU water block and through the Dual Bay Pumps back into the 480 Rad.
But again, that's just my findings. On your next SR-2 build, (if you're using 2 Rads) you might want to try it to see if what I'm saying has the validity that I'm leaning towards. In any case, I'm still OCCT TT'ing... will get back to you shortly, you're info (like Tutor's) has been very helpful and an amazing guide, thank you...
Last edited:
Wait, (and again, please forgive my ignorance), but now I am confused?? I thought you've been advocating not to go above 1.35 V while all the time you have been able to keep that steady GB score (of 40,100 with a Vcore that is lower than 1.35 V), so is that still the case, or are you only reaching those high GB scores when you're using those higher voltages? If that's the case then, what is your current GB score with your current voltages? Thx...
To still get my high scores, I have to do what I did to get them the first time. But I repeat, "Don't go over 1.35 Vcore unless YOU are willing to fully except all of the consequences." That's my disclaimer. My current settings are task based and, thus, vary based on my needs. If the task is light, I set my CPU 0 voltage low, e.g., ~ 1.25V + ~ 157 bclk; if it's moderate, e.g., ~ 1.28V + ~ 163 bclk; if it's high, e.g., ~1.34V + ~ 167 - 173 bclk; if all hell has broken lose, e.g., ~ 1.4V + ~ 183 bclk; and if its the end of time or I've had too much wine, the sky's the limit. But that's just the way I roll the dice.
BTW - That computer in the pic you last posted here is not mine! Mine also has fans underneath.
Last edited:
Hi BDF, well I just started it up the system today to OCCT TT and all I saw was FF, so I shut down and pressed the red Clear CMOS and CMOS black button and then rebooted. I went into my BIOS and loaded up the last saved settings:
Frequency/Voltage Control (main screen)
No Dummy O.C.
Target CPU - calculated value....= cpu multi x CPU freq
CPU Freq. - 160
PCIE Freq. - 100
CPU Multi - 13
QPI - 5.8666
Memory Freq. - 1333 Mhz
Frequency/Voltage Control
CPU 0 Vcore boot - 1.325
CPU 1 Vcore boot - 1.325
CPU 0 Vcore eventual - 1.325
CPU 1 Vcore eventual - 1.325
CPU VTT both boot - 1.25V
CPU VTT both eventual - 1.25V
CPU 0 DIMM Vcore - 1.65V
CPU 1 DIMM Vcore - 1.65V
CPU 0 PLL Voltage - AUTO
CPU 1 PLL Voltage - AUTO
IOH - AUTO
CPU 0 DDR PWM Freq - 800 KHz (AUTO)
CPU 1 DDR PWM Freq - 800 KHz (AUTO)
Memory Configuration
DRAM tCL - 8
DRAM tRCD - 10
DRAM tRP - 8
DRAM tRAS - 26
Command Rate - 1
But this time I entered Vcore - 1.39375 and it still FF'd on me. Then I tried yesterday's settings before it failed (after an 1hr 13min of OCCT TT'ing). That Vcore setting was: 1.3875. It still FF'd on me. Then I went back to Vcore 1.325 and still FF'd. Then just entered in the default and it started. For some reason it won't let me even try now, don't know why. I even disconnected everything (including the battery) and cleared CMOS, waited over 5 min., then reconnected everything and still the same problem... FF. So I don't know what to do at this point... Any thoughts? Thanks...
----------
Great, now I have a better idea of what you're talking about with your settings. Thanks...
Again, I was just messin'...
Frequency/Voltage Control (main screen)
No Dummy O.C.
Target CPU - calculated value....= cpu multi x CPU freq
CPU Freq. - 160
PCIE Freq. - 100
CPU Multi - 13
QPI - 5.8666
Memory Freq. - 1333 Mhz
Frequency/Voltage Control
CPU 0 Vcore boot - 1.325
CPU 1 Vcore boot - 1.325
CPU 0 Vcore eventual - 1.325
CPU 1 Vcore eventual - 1.325
CPU VTT both boot - 1.25V
CPU VTT both eventual - 1.25V
CPU 0 DIMM Vcore - 1.65V
CPU 1 DIMM Vcore - 1.65V
CPU 0 PLL Voltage - AUTO
CPU 1 PLL Voltage - AUTO
IOH - AUTO
CPU 0 DDR PWM Freq - 800 KHz (AUTO)
CPU 1 DDR PWM Freq - 800 KHz (AUTO)
Memory Configuration
DRAM tCL - 8
DRAM tRCD - 10
DRAM tRP - 8
DRAM tRAS - 26
Command Rate - 1
But this time I entered Vcore - 1.39375 and it still FF'd on me. Then I tried yesterday's settings before it failed (after an 1hr 13min of OCCT TT'ing). That Vcore setting was: 1.3875. It still FF'd on me. Then I went back to Vcore 1.325 and still FF'd. Then just entered in the default and it started. For some reason it won't let me even try now, don't know why. I even disconnected everything (including the battery) and cleared CMOS, waited over 5 min., then reconnected everything and still the same problem... FF. So I don't know what to do at this point...
Any thoughts? Thanks...----------
Great, now I have a better idea of what you're talking about with your settings. Thanks...
Again, I was just messin'...
The only time I've hit FF recently is because the first signal tweak, IOH QPI 0 Signal, was set too negative. Right now, if I go below -76, I get FF.
Well, I'm at -75. I went down by 10 to -65 and nothing, then -55 and then -50 and it finally started. I have everything set as I did before. This time my Vcore is now 1.40V. We'll see if this passes the 2 hours OCCT TT.
BTW, what does putting the IOH QPI 0 at -XX do for the system? Thanks...
Last edited:
Yeah, try -60 and see if that helps. I just found what the lowest was that I could hit by changing it by 1, rebooting, seeing if I got FF, and repeating. It will probably change if I change voltages or anything...