Overclock.net › Forums › AMD › AMD CPUs › [OFFICIAL] FX-8320/FX-8350 Vishera Owners Club
New Posts  All Forums:Forum Nav:

[OFFICIAL] FX-8320/FX-8350 Vishera Owners Club - Page 462

post #4611 of 67906
It might not be the right thread since this is an owner's thread but Amazon has the FX-8320 on sale for $166 right now:

http://www.amazon.com/AMD-FX-Series-Eight-Core-Processor-FD8320FRHKBOX/dp/B009O7YU56
    
CPUMotherboardGraphicsRAM
Intel Core i5-4670K Gigabyte GA-Z97-D3H PowerColor Radeon HD7870 Myst (Tahiti LE) Samsung MV-3V4G3 4×4 GB DDR3-1600 
Hard DriveHard DriveCoolingOS
Crucial M4 SSD 256 GB WD Green WD15EARS 1.5 TB Cooler Master Hyper 212 EVO Windows 10 Pro 
MonitorKeyboardPowerCase
Viewsonic VX2268WM 120Hz Redragon K552-N KUMARA Cooler Master Real Power Pro 550W NZXT LEXA 
MouseAudio
HP VoodooDNA Gaming Mouse Samson SR850 Headphones 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
Intel Core i5-4670K Gigabyte GA-Z97-D3H PowerColor Radeon HD7870 Myst (Tahiti LE) Samsung MV-3V4G3 4×4 GB DDR3-1600 
Hard DriveHard DriveCoolingOS
Crucial M4 SSD 256 GB WD Green WD15EARS 1.5 TB Cooler Master Hyper 212 EVO Windows 10 Pro 
MonitorKeyboardPowerCase
Viewsonic VX2268WM 120Hz Redragon K552-N KUMARA Cooler Master Real Power Pro 550W NZXT LEXA 
MouseAudio
HP VoodooDNA Gaming Mouse Samson SR850 Headphones 
  hide details  
Reply
post #4612 of 67906
Quote:
Originally Posted by gertruude View Post

Well i see it as this way. For my 8350 OC of 4.8 at 1.428vcore IBT after 20 runs on very high say its stable and OCCT i can run for as long as i like with no problems. Then i go into prime95 and after awhile one of the cores fail. To get prime to run for as long as the other 2 i need to put more vcore into the chip. Its not by much for me but it does need increasing
End of the day which software would you trust to say its stable. When IBT and OCCT say its stable and i can game all day encode a few videos with no problems at all or do i go with prime95 that says its unstable and bump up the vcore,
For me IBT does run hotter than prime95. IBT and OCCT seem to run at the same temps and prime runs a bit cooler not much but it does runa bit cooler. So the claim by the author of ibt is probably justified at least for me.

As I have said before in this thread, IBT can do 10 passes on my box at 5 GHz on air, when Prime 95 BSODs immediately after startup.
BSODs should never happen with any program if the system is stable. You can't blame any bug in Prime95 for a BSOD. A BSOD is simply proof that the box is unstable. IBT is at best misleading, at worst useless. I fail to see the value in using it.

That said, I'm not ruling out some problem with Prime95.
I did the following run last night :

CPU clock ratio X18.5 4440
Northbridge freq x8 1920
CPB off
CPU host control auto
CPU frequency 240
PCIE clock 100

HT link width auto
HT link frequency x10 2400
DRAM EOCP Disabled
set memory clock Manual
memory clock X6.66 1599

voltage auto
CPU PLL normal 2.5V
DRAM voltage normal 1.5V
DDR VTT normal 0.75V
NB voltage normal 1.10V
HT Link voltage control normal 1.2V
NB/PCIe/PLL Voltage Control normal 1.8V
CPU NB VID normal
CPU voltage +0.05V 1.3875V

Load Line Calibration Extreme

BIOS boot OK
Windows boot OK
IBT

max temp 51°C

Prime 95 failed at 52 minutes


4.44 GHz is not much of an overclock at all, and I'm using a big case with all the fans installed and a fancy air cooler - NH-D14.
I am surprised that Prime95 lasted only this long.

FYI, the "max temp" I'm recording is from the CPU under HWmon.

I'm not sure if I should be recording some other value. The TMPIN2 motherboard sensor, which I believe corresponds to the CPU socket, got a max temp of 62°C.
post #4613 of 67906
Quote:
Originally Posted by madbrain View Post

As I have said before in this thread, IBT can do 10 passes on my box at 5 GHz on air, when Prime 95 BSODs immediately after startup.
BSODs should never happen with any program if the system is stable. You can't blame any bug in Prime95 for a BSOD. A BSOD is simply proof that the box is unstable. IBT is at best misleading, at worst useless. I fail to see the value in using it.
That said, I'm not ruling out some problem with Prime95.
I did the following run last night :
CPU clock ratio X18.5 4440
Northbridge freq x8 1920
CPB off
CPU host control auto
CPU frequency 240
PCIE clock 100

HT link width auto
HT link frequency x10 2400
DRAM EOCP Disabled
set memory clock Manual
memory clock X6.66 1599

voltage auto
CPU PLL normal 2.5V
DRAM voltage normal 1.5V
DDR VTT normal 0.75V
NB voltage normal 1.10V
HT Link voltage control normal 1.2V
NB/PCIe/PLL Voltage Control normal 1.8V
CPU NB VID normal
CPU voltage +0.05V 1.3875V

Load Line Calibration Extreme

BIOS boot OK
Windows boot OK
IBT

max temp 51°C

Prime 95 failed at 52 minutes
4.44 GHz is not much of an overclock at all, and I'm using a big case with all the fans installed and a fancy air cooler - NH-D14.
I am surprised that Prime95 lasted only this long.
FYI, the "max temp" I'm recording is from the CPU under HWmon.
I'm not sure if I should be recording some other value. The TMPIN2 motherboard sensor, which I believe corresponds to the CPU socket, got a max temp of 62°C.


sounds like your chip cannot do -ghz, my two cents. need more voltage.
Edited by pwnzilla61 - 11/30/12 at 5:00pm
Rigamormus
(19 items)
 
  
CPUMotherboardGraphicsRAM
AMD FX-8350@5Ghz Asus Crosshair V Formula MSI GTX 680 Lightning@1333+600mem 8gb Gskill Sniper 1600 9-9-9-24 
Hard DriveHard DriveHard DriveOptical Drive
2x500gb Crucial M4 OCZ Vertex III walmart brand 
CoolingCoolingOSMonitor
Raystorm custom loop. Coolgate 360 Windows 7 64 Ultimate LG IPS236v 
KeyboardPowerCaseMouse
Razer Black Widow Ultimate 2013 Corsair TX850M Corsair 600t SE Razer Deathadder 
Mouse PadAudioAudio
OCZ Regulator Razer Tiamat 2.2 Creative Sound Blaster Recon3D/audigy 
  hide details  
Reply
Rigamormus
(19 items)
 
  
CPUMotherboardGraphicsRAM
AMD FX-8350@5Ghz Asus Crosshair V Formula MSI GTX 680 Lightning@1333+600mem 8gb Gskill Sniper 1600 9-9-9-24 
Hard DriveHard DriveHard DriveOptical Drive
2x500gb Crucial M4 OCZ Vertex III walmart brand 
CoolingCoolingOSMonitor
Raystorm custom loop. Coolgate 360 Windows 7 64 Ultimate LG IPS236v 
KeyboardPowerCaseMouse
Razer Black Widow Ultimate 2013 Corsair TX850M Corsair 600t SE Razer Deathadder 
Mouse PadAudioAudio
OCZ Regulator Razer Tiamat 2.2 Creative Sound Blaster Recon3D/audigy 
  hide details  
Reply
post #4614 of 67906
Quote:
Originally Posted by bios0110 View Post

madbrain, it is also my opinion that Prime95 is much more useful. The reason I told you to run ITB is only to generate heat. That seems to be the common consensus (that it heats well). It is not as good as Prime95 at catching stability issues (as you have said). This is why I said to run Prime95 after IBT (or any heat gen soft).
Personally I run Prime95 FFT to make heat (this test CPU alone for stability). Causes enough heat. Then if that is stable blend with Prime95 (this also tests ram).
There is always some psychology with overclocking. Some people are so desperate for an extra few hundred Mhz that they will subconsciously avoid tests that they know they will fail. Other feel that the system isn't stable unless it runs at full throttle for one week. To each his own. I think that Prime95 for 12hrs is the minimum proof for stability.

Thanks.
Yes, when I have OC'ed in the past, I always considered Prime 95 from 12-24 hours to be stable.

In fact I'm typing this message on a Phenom II x6 1055T box OC'ed to 3.78 GHz (14x270) on air in a very tight and very full 12 year old Lian-li PC-60 case ... It has a GA-890GX-UD3H mobo. Also on air, with Hyper 212 cooler. And the biggest fan in the case is 80mm. There are 4 optical drives, one floppy drive (no laughing allowed), card reader, 2 HD, one SSD. And I think almost all the slots are full. And this one is Prime95 stable 24 hours, at least it was back in the spring when I setup the OC.

So ... this FX-8350 in the much bigger HAF-XM case with 1200W PS, only half as many drives, and so many more larger fans, is frustrating me more than a little bit. with the OC . Prime 95 sure doesn't seem to like it.
I have just reverted everything to stock right now, except for the APM, C1E, C6 and CnQ. And going to see if Prime95 lasts 12 hours on it.
post #4615 of 67906
Quote:
Originally Posted by anubis44 View Post

Although I'm still testing out overclocks with my own new FX-8350, I would already suggest that you might want to use the 'High' LLC setting, as opposed to 'Very High' or 'Extreme'. The reason is that the 'High' setting, at least on Gigabyte boards, will stabilize the voltage at your requested level, but no higher than that, and this is probably the best setting for a cooling-limited situation, especially air-cooling. You air cooler is already being strained to its very limits with the 4.6GHz overclock, so that core failure is probably happening due to a very slight increase in heat build-up that's pushing the core just past its thread failure temp (perhaps the ambient room temp went up a degree or so, who knows). I think the 'Very High' and 'Extreme' settings might work better for those using good water cooling, and who have a little bit of extra cooling margin available if the voltage ramps up slightly under load.
Just my thoughts.

Thanks, I will try the LLC high setting next.

Re: the failures I have seen with P95, I thought they were due to high temps at first. But I also seen them happen even in runs when the peak temp was not that high. I wasn't in front of the box to witness when failures happened so I can't say if they coincide with the high temp or not.
Most of the time at high clocks though, I see the failures in prime95 happen almost right away, before the temps have had time to build up.

Because I'm on air I have to keep my voltages relatively low, there is no point trying to get my voltage to 1.5V as I know the heat will build up too much for a 12 hour run for sure...
post #4616 of 67906
Quote:
Originally Posted by Solders18 View Post

Lol stable enough... if you aren't stable at 4.6 what makes you think you will be stable at 4.8? I don't mean to be snide but stability is either a yes or no, no kinda's.

What I meant is that I'm never going to run something at 100% for more than 6 hours in the real world.

So most likely my 4.6 OC was sufficiently stable for my use. But it wouldn't be if I was doing number crunching 24/7.

I would prefer to have a 12-24 hour stable prime95 stable test of course, though.
post #4617 of 67906
I hear you brother, I use NCIX all the time and they will match prices. I can call them and when they have it all together I drop in. No hassles, great staff, {Langley, B.C.}
post #4618 of 67906
playing around with my two rigs.

i5 3570k + GTX 680 (998USD) vs. FX8320 + GTX 670 (750USD)

i5 3570k + gtx 680
i5 3570k + GTX 680

FX8320 + GTX 670 (would've been better if I had a 7970 which is actually cheaper than a gtx 670 from where I'm from)
FX-8320 + GTX 670

25% cheaper but only 10% minus the performance. not bad eh biggrin.gif

more coming up
post #4619 of 67906
Quote:
Originally Posted by bios_R_us View Post

Hi everyone,
After updating to the latest version of HWINFO32, I've found it reporting a new thermal sensor, on the mobo (I'm using the 990FXA-UD3). It's called CPU and it seems a lot more accurate than either the socket temp or the core temp, in that it is very close to the core temp under load (usually the same or 0.5C difference) but does not drop to insane values like 12C in idle, but stays somewhere at a believable 35C. This would make the sensor very valuable if proven to be accurate I guess, but more of you guys can try if out if you want and let me know what you think. I have no idea what it does on other boards :-)
Here's a screenshot:

RED is the Core temp as reported by the CPU sensor
BLUE is the socket temp that I've been reading so far on the mobo
GREEN is the new temp that was not present in any monitoring program so far.. and that seems really accurate.

I downloaded hwinfo64 and tested it on my 990FXA-UD3 rev 1.1 .
I like the program, but I don't really see anything new.

The green CPU sensor you outlined is the same as what I see in CPUID Hwmon under "TMPIN1" .
Just under a different name.

The blue one is TMPIN2 .

The red one is the same as the CPU temp in hwmon also, except the version of HWmon I have (build 135) shows separate temp for each core, and HWinfo64 just shows one temp.
For the most part when running stress/stability tests, the individual core temps are very close, but sometimes they are off by about 1°C.
I believe later builds of HWmon was changed to show just one CPU temp instead of per-core.
post #4620 of 67906
Quote:
Originally Posted by pwnzilla61 View Post

Here is my 24/7 clock for 5ghz if any is interested.
The only thing I did was disable, all power saving features, all llc to max expect cpu llc to ultra. The only voltage I changed was cpu.

What kind of cooling do you use ? Does prime95 run 24/7 also?
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: AMD CPUs
Overclock.net › Forums › AMD › AMD CPUs › [OFFICIAL] FX-8320/FX-8350 Vishera Owners Club