Message boards :
Number crunching :
Getting the most Seti cpu production out of your Amd 7 2600/2700 cpu
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5 · 6 · Next
Author | Message |
---|---|
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
I also got all the M.2 slots working on the C7H. But the same post failure # "90" or "06" which way is up? I think it is 90. Since it seems to be stalling on the boot disk? I tried moving the HD data cable around. No luck. I will look. TY. Tom A proud member of the OFA (Old Farts Association). |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
I would think it would lead to instability. The current Ryzen 3000 BIOS doesn't even qualify as an "alpha" let alone a "beta" Whispering...... got the new, gigantic air cpu cooler on. After a some minutes getting the gpus all happy again (I removed the #2 PSU/harness for later deployment). And confirming the temps were reasonably stable/low at 3.7GHz I jumped it to 4.0Ghz. Normally it will crash withing 3-5 minutes. Since it is still running. I am presuming I am stable. I installed the llc 3, phase control optimize, current delivery 120% parameters. May need to try w/o them because the temp has gone up some..... Its at 74C right now, and at 3.7Ghz it passed 85+C regularly with the stock cpu cooler from Amd. Anyway, lets see what happens to the RAC overnight. It had flattened out again. But at least it wasn't dropping. Tom A proud member of the OFA (Old Farts Association). |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
What is your Vcore at 4Ghz loading? You should be somewhere around 1.35V or a bit lower. I end up around 1.33- 1.34V on Auto at 4.025Ghz. LLC on Auto. Power delivery at 120%. On 280mm AIO cooling though compared to your air cooler. On air, you need to run LLC2 or LLC3 probably. The LLC is what forces the temps up as it doesn't let the Vcore voltage droop as much and the power usage is greater. But if the LLC gives you more stability without dropping the host offline, then you just deal with the sligher hotter temps and don't go as far with the core clocks. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
What is your Vcore at 4Ghz loading? You should be somewhere around 1.35V or a bit lower. I end up around 1.33- 1.34V on Auto at 4.025Ghz. LLC on Auto. Power delivery at 120%. On 280mm AIO cooling though compared to your air cooler. On air, you need to run LLC2 or LLC3 probably. The LLC is what forces the temps up as it doesn't let the Vcore voltage droop as much and the power usage is greater. But if the LLC gives you more stability without dropping the host offline, then you just deal with the sligher hotter temps and don't go as far with the core clocks. Not sure what it was. It crashed during the night. I love the cmos clear button. Everything but the multiplier are auto. Plus preset for D.O.....? Now have it at 3.7Ghz. Lets see now..... tom@EJS-GIFT:~$ sensors asuswmisensors-isa-0000 Adapter: ISA adapter CPU Core Voltage: +1.27 V CPU SOC Voltage: +1.12 V DRAM Voltage: +1.35 V VDDP Voltage: +0.28 V 1.8V PLL Voltage: +1.85 V +12V Voltage: +12.04 V +5V Voltage: +5.01 V 3VSB Voltage: +3.36 V VBAT Voltage: +3.23 V AVCC3 Voltage: +3.36 V SB 1.05V Voltage: +1.07 V CPU Core Voltage: +1.27 V CPU SOC Voltage: +1.13 V DRAM Voltage: +1.36 V CPU Fan: 1457 RPM Chassis Fan 1: 0 RPM Chassis Fan 2: 0 RPM Chassis Fan 3: 0 RPM HAMP Fan: 0 RPM Water Pump: 0 RPM CPU OPT: 1450 RPM Water Flow: 0 RPM AIO Pump: 0 RPM CPU Temperature: +70.0°C CPU Socket Temperature: +53.0°C Motherboard Temperature: +34.0°C Chipset Temperature: +51.0°C Tsensor 1 Temperature: +216.0°C CPU VRM Temperature: +56.0°C Water In: +216.0°C Water Out: +216.0°C CPU VRM Output Current: +84.00 A k10temp-pci-00c3 Adapter: PCI adapter temp1: +71.1°C (high = +70.0°C) asus-isa-0000 Adapter: ISA adapter cpu_fan: 0 RPM tom@EJS-GIFT:~$ A proud member of the OFA (Old Farts Association). |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
1.27V is fine for 3.7Ghz. Question was what was/is the voltage at 4Ghz. Unless you have the crappiest cpu ever produced, a 2700 should be able to run 4Ghz. Again the cpu voltage needs to be around 1.35V. I see your memory running at XMP voltage. Are you running them stock? Why don't you bump them up to 1.365 -1.37V. Do you know why you are crashing? Have you looked at the logs to see what the error is? Do you ever run a stress tester at your new attempted clock frequency to see if your settings are stable? A half hour of Prime95 small FFT's will quickly show errors if you don't have enough voltage. A half hour run of stressapptest will prove that your memory is stable at the attempted clocks. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
1.27V is fine for 3.7Ghz. Question was what was/is the voltage at 4Ghz. Unless you have the crappiest cpu ever produced, a 2700 should be able to run 4Ghz. Again the cpu voltage needs to be around 1.35V. Ran all day at 3.7Ghz. Got home from work. Had a system update (security only and core image). Booted it back up to 4.0 Ghz. Without anything accept the D.O.C.P. Standard on the ram. Everything else is auto. Current (pun?) results look like this: tom@EJS-GIFT:~$ sensors asuswmisensors-isa-0000 Adapter: ISA adapter CPU Core Voltage: +1.30 V CPU SOC Voltage: +1.12 V DRAM Voltage: +1.35 V VDDP Voltage: +0.44 V 1.8V PLL Voltage: +1.96 V +12V Voltage: +11.99 V +5V Voltage: +4.99 V 3VSB Voltage: +3.36 V VBAT Voltage: +3.25 V AVCC3 Voltage: +3.36 V SB 1.05V Voltage: +1.07 V CPU Core Voltage: +1.30 V CPU SOC Voltage: +1.13 V DRAM Voltage: +1.36 V CPU Fan: 1464 RPM Chassis Fan 1: 0 RPM Chassis Fan 2: 0 RPM Chassis Fan 3: 0 RPM HAMP Fan: 0 RPM Water Pump: 0 RPM CPU OPT: 1458 RPM Water Flow: 0 RPM AIO Pump: 0 RPM CPU Temperature: +76.0°C CPU Socket Temperature: +56.0°C Motherboard Temperature: +35.0°C Chipset Temperature: +52.0°C Tsensor 1 Temperature: +216.0°C CPU VRM Temperature: +58.0°C Water In: +216.0°C Water Out: +216.0°C CPU VRM Output Current: +93.00 A k10temp-pci-00c3 Adapter: PCI adapter temp1: +76.1°C (high = +70.0°C) asus-isa-0000 Adapter: ISA adapter cpu_fan: 0 RPM tom@EJS-GIFT:~$
I had limited time this morning so instead of relying on some of the other bios presets I used the DOCP Standard which runs ram at 3200 with all the other stuff that appears to be "stock" overclock.
The only log I look at regularly is erased when Boinc is restarted. I will be researching the why as I dig out the other Linux info that has moldered/disappeared in my head. The OS did throw a report to developer error that I submitted this morning. But since there was a core image update I went ahead and updated that.
I have regularly had Seti crash things when I over do the OC sets so until now I haven't seen a need. With WOW under way, I am reluctant to take the time when I don't have them already setup and had practice using those stress testers. I am going to babysit the machine at 4.0Ghz all evening. If it doesn't crash by bedtime I am hoping it will run overnight. Meanwhile back to the log research. Tom A proud member of the OFA (Old Farts Association). |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
1.30V isn't enough for BOINC load at 4Ghz. If Auto is setting 1.30V for the 4 gig VID values, you are going to have to use an Offset for Vcore instead of Auto. Change to Offset for Cpu Voltage and input 35mV to 50mV of additional voltage. As I have said previously, you need to see Vcore under BOINC load of around 1.33 - 1.35V for 4Ghz on the C7H and 1002 BIOS. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
1.30V isn't enough for BOINC load at 4Ghz. If Auto is setting 1.30V for the 4 gig VID values, you are going to have to use an Offset for Vcore instead of Auto. Change to Offset for Cpu Voltage and input 35mV to 50mV of additional voltage. As I have said previously, you need to see Vcore under BOINC load of around 1.33 - 1.35V for 4Ghz on the C7H and 1002 BIOS. I just tried the manual overide with LLC 3 (from a previous change) and got sensors asuswmisensors-isa-0000 Adapter: ISA adapter CPU Core Voltage: +1.27 V CPU SOC Voltage: +1.12 V DRAM Voltage: +1.35 V There is something hinky about doing "offset" s. I haven't been able to get a number with or without a decimal to enter. Tom A proud member of the OFA (Old Farts Association). |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
1.30V isn't enough for BOINC load at 4Ghz. If Auto is setting 1.30V for the 4 gig VID values, you are going to have to use an Offset for Vcore instead of Auto. Change to Offset for Cpu Voltage and input 35mV to 50mV of additional voltage. As I have said previously, you need to see Vcore under BOINC load of around 1.33 - 1.35V for 4Ghz on the C7H and 1002 BIOS. The hinky is you have to use the + and - keys can't just put in a number. :) Here is with an offset of 35mm and all the Digi-Vrm settings. tom@EJS-GIFT:~$ sensors asuswmisensors-isa-0000 Adapter: ISA adapter CPU Core Voltage: +1.32 V CPU SOC Voltage: +1.12 V DRAM Voltage: +1.35 V VDDP Voltage: +0.44 V 1.8V PLL Voltage: +1.96 V +12V Voltage: +12.04 V +5V Voltage: +4.99 V 3VSB Voltage: +3.36 V VBAT Voltage: +3.25 V AVCC3 Voltage: +3.38 V SB 1.05V Voltage: +1.07 V CPU Core Voltage: +1.32 V CPU SOC Voltage: +1.13 V DRAM Voltage: +1.36 V CPU Fan: 1470 RPM Chassis Fan 1: 0 RPM Chassis Fan 2: 0 RPM Chassis Fan 3: 0 RPM HAMP Fan: 0 RPM Water Pump: 0 RPM CPU OPT: 1466 RPM Water Flow: 0 RPM AIO Pump: 0 RPM CPU Temperature: +74.0°C CPU Socket Temperature: +52.0°C Motherboard Temperature: +34.0°C Chipset Temperature: +52.0°C Tsensor 1 Temperature: +216.0°C CPU VRM Temperature: +52.0°C Water In: +216.0°C Water Out: +216.0°C CPU VRM Output Current: +82.00 A asus-isa-0000 Adapter: ISA adapter cpu_fan: 0 RPM k10temp-pci-00c3 Adapter: PCI adapter temp1: +74.6°C (high = +70.0°C) tom@EJS-GIFT:~$ ^C tom@EJS-GIFT:~$ I am going to reset everything, go back to DOCP ram settings, 40 cpu multipler and see what happens to the volts and heat. If it will run LLC auto it should be cooler (I read someplace :^) Tom A proud member of the OFA (Old Farts Association). |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
With the offset at 0.35~ I got this: tom@EJS-GIFT:~$ sensors asuswmisensors-isa-0000 Adapter: ISA adapter CPU Core Voltage: +1.25 V CPU SOC Voltage: +1.12 V DRAM Voltage: +1.35 V So the LLC 3 and maybe other stuff made a voltage difference. And it is a tiny bit cooler. The highest cpu voltage offset it will let me do is 0.45 so will try that next. Tom A proud member of the OFA (Old Farts Association). |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
tom@EJS-GIFT:~$ sudo modprobe asus-wmi-sensors [sudo] password for tom: tom@EJS-GIFT:~$ sensors asuswmisensors-isa-0000 Adapter: ISA adapter CPU Core Voltage: +1.36 V CPU SOC Voltage: +1.12 V DRAM Voltage: +1.35 V So lets try 0.41 A proud member of the OFA (Old Farts Association). |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
You change from Auto to Offset. Don't change "sign" It should default to "+" then move to Voltage Override and just keep hitting the "plus" key on the ten key keypad to keep adding 0.00625V to the offset until you to total up 0.03125V or in other words 31 mV to the 1.30V to get to 1.33V or keep adding. The smallest voltage change possible is 0.00625V because that is the LSB value in the D to A conversion. Save BIOS and reboot and check your readback voltage which should now be 1.33V. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
What BIOS are you on? Are you on 2501 BIOS by some chance? I gave instructions for the most stable 1002 BIOS. BIOS 2501 is totally screwed up with Auto voltages when moving multiplier off Auto and setting a manual multiplier. It defaults to 1.10V for VID for all multipliers. There should not be any limit in the amount of offset added. You can add a full volt if you want to. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
You change from Auto to Offset. Don't change "sign" It should default to "+" then move to Voltage Override and just keep hitting the "plus" key on the ten key keypad to keep adding 0.00625V to the offset until you to total up 0.03125V or in other words 31 mV to the 1.30V to get to 1.33V or keep adding. The smallest voltage change possible is 0.00625V because that is the LSB value in the D to A conversion. At 0.40625~ I am getting tom@EJS-GIFT:~$ sensors asuswmisensors-isa-0000 Adapter: ISA adapter CPU Core Voltage: +1.32 V CPU SOC Voltage: +1.12 V DRAM Voltage: +1.35 V VDDP Voltage: +0.45 V 1.8V PLL Voltage: +1.96 V +12V Voltage: +11.99 V +5V Voltage: +4.99 V 3VSB Voltage: +3.36 V VBAT Voltage: +3.25 V AVCC3 Voltage: +3.36 V SB 1.05V Voltage: +1.07 V CPU Core Voltage: +1.33 V CPU SOC Voltage: +1.13 V DRAM Voltage: +1.37 V CPU Fan: 1457 RPM Chassis Fan 1: 0 RPM Chassis Fan 2: 0 RPM Chassis Fan 3: 0 RPM HAMP Fan: 0 RPM Water Pump: 0 RPM CPU OPT: 1457 RPM Water Flow: 0 RPM AIO Pump: 0 RPM CPU Temperature: +77.0°C CPU Socket Temperature: +55.0°C Motherboard Temperature: +34.0°C Chipset Temperature: +51.0°C Tsensor 1 Temperature: +216.0°C CPU VRM Temperature: +56.0°C Water In: +216.0°C Water Out: +216.0°C CPU VRM Output Current: +93.00 A k10temp-pci-00c3 Adapter: PCI adapter temp1: +76.5°C (high = +70.0°C) asus-isa-0000 Adapter: ISA adapter cpu_fan: 0 RPM A proud member of the OFA (Old Farts Association). |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
Try to run overnight on that 1.32V and see if it holds up. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
Try to run overnight on that 1.32V and see if it holds up. Since I am tired of screwing around with it and it is still running nearly an hour later..... its a wrap (or is it a rack)? Collapse (me not the computer). Tom A proud member of the OFA (Old Farts Association). |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
I was looking again at your sensors readout. I don't think you got around to installing the socket fan I talked about have you? That would help stability a lot. Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Tom M Send message Joined: 28 Nov 02 Posts: 5124 Credit: 276,046,078 RAC: 462 |
I was looking again at your sensors readout. I don't think you got around to installing the socket fan I talked about have you? That would help stability a lot. It turns out I will need to buy the suggested fan. The ones I have available are large enough I was having trouble installing it. I have a bar on the rack that limits how far over I can install the fan. I will go back and re-read that part of the thread then order a couple three. I expect to run Amd cpus for years so I want them a cool as I can. It ran all night. !!! At 4Ghz !!!!!!!! Tom ps. Yes I am still thinking about a 2950x..... "more cores...." ;) A proud member of the OFA (Old Farts Association). |
Keith Myers Send message Joined: 29 Apr 01 Posts: 13164 Credit: 1,160,866,277 RAC: 1,873 |
I looked at your overnight results. You are overcommitted on the cpu. So how many cpu tasks each are you running with how many gpu tasks each? Think this host is the six gpu card host, correct? Seti@Home classic workunits:20,676 CPU time:74,226 hours A proud member of the OFA (Old Farts Association) |
Ian&Steve C. Send message Joined: 28 Sep 99 Posts: 4267 Credit: 1,282,604,591 RAC: 6,640 |
if he keeps trying to use 0.1 CPU per GPU in his settings while setting 90% CPU use, he will keep pegging his CPU out since BOINC "thinks" he's only using that much, when he's really using 1:1 CPU:GPU with -nobs. for example, 6 GPUs, 16 threads, -nobs, 0.1:1, 90% CPU setting: 6/16 used for GPU = 62% CPU Actually free Boinc thinks you are only using <1 thread (0.6) and 93% free Boinc will spin up 14 CPU tasks You now have 20 WUs running on a 16 thread CPU = 100% use, pegged or, 6 GPUs, 16 threads, -nobs, 1:1, 90% CPU setting: 6/16 used for GPU = 62% CPU Actually free Boinc now aware of your GPU impact to CPU and correctly knows only 62% free Boinc will spin up only 8 extra CPU jobs You now have 14 WUs running, and your CPU use percentage will be 90% as you've intended or you can do some hacky kludgey settings to achieve the same thing in a round about way with no benefit and more confusing configurations 6 GPUs, 16 threads, -nobs, 0.1:1, 50-55% CPU setting: 6/16 used for GPU = 62% CPU Actually free Boinc thinks you are only using <1 thread (0.6) and 93% free Boinc will spin up 8 CPU tasks You now have 14 WUs running on a 16 thread CPU = 90% which do you think makes more sense? Seti@Home classic workunits: 29,492 CPU time: 134,419 hours |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.