i'll have to take a look inside the case to see how big the fan can be and account for mounting parameters and also if the air will go straight to the back of the mobo. i'm not 100% on this and don't want to crack the case at this second. going to run prime95 for two hours and see how things are at that point.
...Okay guys let us remember that this is an FX-9590 and not an FX-8350. We have had what now about 6 or so FX-9590s come in for help and maybe 3 or so FX-9370?. Many of those were in the AMD mobo section and the 'biggest' task was getting cooled enough to be stable once the multiplier was manually set and Green turned off making ALL 8 cores run at the same time. So we need to diagnose carefully. READ what is written I find to be both very difficult and also best for troubleshooting. I have to go back and read at least 10 posts in arrears to get current for last post. I am a little slow you know!!!
So what is it we might really know?
1. > Memory: x2 Corsair Dominator CMD16GX33M2A1866C10 8GB 1866Mhz < Okay remember here in the last week or so when trying to help "Golodkin" and his board and Ram questioning? The biggest contributor to Ram issues has been Kingston, but there have been problems with other 8gig sticks of ram. Corsair being one that bites at times. What Ram is in play here? I could stand to leave the Ram at DDR3-1333 for the time being.
2. > Okay Samena is not really giving enough information to forge forward. 24 multiplier crashed the WHOLE system according to his words. But it would be nice to see the first 1 minute and 30 seconds before the crash to try and get an eye on what is happening. Without as close to time of FAILURE images of what is happening...well we still must speulate.
3. > So far Samena's setup is exhibiting just about what all the other big FX-9xxx cpus have shown us. Not even really stable at stock non-TurboCore speed. Just barely able to run the numbers and the least change equals failure one way or the other.
4. > If he (Samena) thinks the cpu is locked then he does not have the board in manual mode??? Somebody with Saberkitty R2.0 needs to put up some pics of what needs setting to go manual for full user control.
5. > We need to get this sort of straight about varying cpu speeds. Cool N Quiet does not by nature drop or raise cpu speed by +/- 300 or so MHz. C N Q drops the cpu speed by a ton from maybe 4.0GHz to the lowest multiplier which is a 7 or 1400Mhz or so. Minor speed changes are NOT indicative of C N Q being enabled. Those minor speed variations are from throttling of some sort. Need to get that ingrained. It gets us chasing the wrong animal thru the woods.
6. > In relation to #5 above, we do not know if Samena has set Windows power manager to "Performance" Mode or not. If so then C N Q also will not work and make the huge drops in cpu speed associated with C N Q kicking when not under load. C N Q does not work when loaded.
7. > I suspect the minor variance in cpu speed to result from the throttling that S_B speaks of seeing with HIS FX-9590 if lhis Socket temp reaches 52c.
8. > EVERY last one of these bigger FX-9xxx cpus we have dealt with has had water cooling of some sort on them but one. Each of those on water "even on good boards" experienced some crazy cpu speeds because the W/Cooler took away air flow over the VRMs and socket area. "Johan" attempted to speak to this in mentioning getting air flow to the 'rear' of the mobo and is and excellent idea. He also mentioned *active cooling* onto/over the VRM section of the mobo and so far every user of water on these FX cpus has had to get air onto the VRMs since air spill-over from a air cooled cpu cooler is GONE with installation of water.
9. > PSU: ANTEC 750W <<< Tells me zero about the power supply since it is not a model number so we can go and check for multiple rails or single rail power supply. System shutting completely OFF as indicated by Samena maybe pointing a finger at LOW power but we need full information in order to try and help him from 1,000s of miles away.
10. > So where the h*ll are we? A. Bound by not enough information. B. Part number information. C. Windows settings suggested being in 'place' and in use. D. Where in the heck are the fans located in the case and direction of air flow. E. If the COOL thing is not addressed first, then I expect Samena to continue to suffer his general inability to make much of any change even if gets board in full on manual mode for user input and save of the input.
11. > We can say power supply might be multi-railed and not have enough power on each +12V rail to handle PUSHED FX-9590. We might say the VRMs are not cooled enough and are throttling. Plenty of stuff to "guess" at but so far I have not seen any of the dearth of information I would need to zero in on what to do in an orderly fashion.
12. > I am going to state a "suspicion" nagging at me. Samena has a diamond in the rough piece of silicon, but without managing his c00ling he is likely to not experience such goodness. But then my suspicion can be flawed by him being at 4.7GHz with lower core temps and is just poised on the next STEP of Vcore upping that will move the temp of core way on up the hill. Not enough information and images and captures to really know.
Need more and more specific information. Period to my way of troubleshooting learned over the course of 29 years of mechanicing. YMMV. Ole faht out.
RGone...skisterzzzzzzzzzz
EDIT:
Johan, Bassnut, Mandrake, ShrimpBrime and most likely myself could have this situation sorted in about 1.5 hours IF the unit were in front of us and us at the keyboard and reading the 'screen'. That is why it is so very important to have a lot of information in picture form to let us see what is going on. This added for information needed emphasis only.
END EDIT.
Sure and post some pics after so we can see how things are going. The fans we described really do help, most of us who hang here every day have done the same to help control the heat these FX CPUs can put out. I have 2 different 9370s and they can be a challenge. I went as far as drilling holes in my case to blow fresh air directly onto the back of the motherboard.
If you cannot control temps at stock, this cpu wont overclock.
Stock all, cpu multi only, leave cool and quiet On voltage to auto and turbo off. Try 4.8ghz. Then 4.9ghz.
Or if manually set voltage oc only, Cpu wont last 2 hours without 1.5250v anything above 4.2ghz.
If FX-9590 throttles, its prone to hang cause of low voltage while manually overclockeded and manually set cpu core voltage.
You need heavy VRM cooling and a stout PSU. This chip pulls 220w np at stock. Manually set voltage green off is a constant strain on the power delivery system. This causes on load the board and or psu to simply shut off. This also can happen while cpu temp spikes faster than the software can pull it. A simple 80c spike is likely to shut the board down. If does not power up, reset cmos for sure.
As for the motherboard and RAM, you should have gone big. The sig board I use is all set up for this monster and fast memory speeds. I never run less that 1067 effective mhz. 11-13-13 is what the board defaults to so it makes ram speeds easy to come by.
In the very very end of the discussion, Shrimp may mention a chiller mod. I has a window unit air conditioner, let me say it helps a lot. Over 5 ghz manually set up BECOMES able to be achieved while core temps and cpu socket temps stay into the lower 40c range with good results and use of 1.6v cpu core and beyond.
I'll post some pics tonight when I get home of how I have my fans set up....... very simple actually.
As far as Corsair ram playing nicely with AMD I am like the others and will recommend GSKILL, that being said I recently acquired a set of Corsair Dominator Platinum 2133 MHZ 8 10 10 27 8Gig (2X4gig) that is playing very nice with my FX rig. I have not had a ton of time lately and my windows is messed up but have had them marginally stable down to 7 10 9 25. But keep in mind this is 1 case....... maybe I can pop in my 1866 set to see how they play as well but again this is just an 8 gig kit too.
Ok so I am checking on on the p95 2hour stability iwth stock settings +disabled features (including:C1/E,C6,CnQ, APM, TURBOCORE).
i did not run test during safe mode. I did however reboot and make sure my bios settings were mostly stock, just absolutely no OC profile at all. All AI Tweak page is stock. I did see some messages popup from HWMonitor saying there were temperatures up to 100+C on mobo and cpu,but not core. Not sure if you can see the Vcore temp, but it's sitting at around 61C at around 11:55am.
-I am only changing 1 setting in AI Tweaker page in bios to set to manual(i forget what the name is but it's the first setting on the AI Tweak which i set to manual when trying to oc. will check and get back to you). i haven't set any jumpers or anything(under the impression it's not needed anymore).
-Power options- i am using the search function (windows+s) and searching power management and selecting power options then selecting high performance from that list. is this what you mean by power manager?
-PSU- I had to look this up at the store site and am pretty positive it is now up to date in my sig. There were only 2 to choose from once i narrowed specs down. BTW what does voltage green mean?
-Fans: 2 in front are in take. 2 on top are from water cooler and are intake. Rear fan is only exhaust
- - - Updated - - -
if someone wants to get an instant messenger then I think it would help more. i'm not sure what is needed for this process at this point. fans;yes, psu;maybe?
I have to ask since I see some odd readings like 2.5v for core and you mentioned a popup warning. Do you have AISuite installed or any other monitoring software? This can cause strange reading with monitoring software when they're both trying to poll the same sensor.
OK do yourself a favour and uninstall it for the time being and any other monitor software except for HWMonitor. I see voltage fluctuations that aren't consistent with the options you disabled in BIOS. So something is causing interference. Also in BIOS enable HPC. Then do a quick run of a few minutes with P95 to see if things are more consistent.
OK that does look better than it did, voltages seem to be more consistent. The next step for me would be to see how it'll do with the voltages and Digi power options set manually. It looks to me like that chip will run at stock with less than 1.45v for the core. So I would set the V+Core manually at 1.45v then in the Digi section set the CPU_LLC to ultra high, CPU_NB to high, and set the current capability for both to 120%.
Then restart and try P95 again for 20 minutes and post results.
You'll have to put the CPU and NB voltage control to manual and then set the CPU Manual Voltage to 1.45 and leave the CPU_NB on auto for the time being.
i want to be clear before i change anything. the text after the '=' is what i see in bios as closest thing. these are the top two options...
CPU_LLC = Load line Calibration?
CPu_NB = LLC/NB?
Current Capability= DRAM Current capability?
All good up to the DRam. There should be an option for both the CPU and NB for current capability. You can likely keep them on auto for now if you feel more comfortable doing that. It likely won't be an issue unless you're pushing for much higher clocks.
Good luck, should go fine but you never know once you take the reigns and the board isn't allowed to make it's own adjustments sometimes things can come off the cracks.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
. will start p95 now.
