Overclock.net › Forums › Graphics Cards › NVIDIA › Any point of overclocking a GTX570 with only 3.2 ghz q9550?
New Posts  All Forums:Forum Nav:

Any point of overclocking a GTX570 with only 3.2 ghz q9550? - Page 2

post #11 of 21
Quote:
Originally Posted by Chris Ihao View Post
No worries mate. I got my answers Just a question regarding that cpu oc. Have you raised the NB voltages properly? Worth trying to raise that one a notch, if you havent already tried it.

Thread is yours;
ok, thanks

hmm, not sure about the options with my mobo. I mean I changed one of the voltages, but, I don't have a ton of options, mostly fsb, multipliers for cpu, ram, maybe 5 voltage options and that's it

I used to get it to 3Ghz easy. I did a 14hr prime run when I built this rig(3 year old now ) on the same settings(as I recall) and temps were at 65C back then and it was rock solid(ram at 833mhz), so, not sure what changed. I'll have to take a second look at the other 4 options, but, the temps are pretty high as is. Would it be wise to raise the voltages more?

Edit: Found pics:
http://www.xbitlabs.com/articles/mai...l_6.html#sect0
There are also FSB, DDR2, PCI-E and MCH(?) aside from the system voltage which I have at 1.650 or so.

Edit2: Here's the odd part:
Running 2 instances of orthos, now 21+ min testing cpu with gromacs code. No issues, errors, etc.... at the same voltage and everything that caused the error before at 18min in. The voltage according to cpuz is fluctuating from 1.184V-1.250V or so.
Edited by BigFan - 1/4/11 at 2:41pm
    
CPUMotherboardGraphicsRAM
Intel Q6600 2.40GHZ Gigabyte GA-P35-DS3L MSI GTX570 @ stock Corsair 2GB DDR2 800MHz 
Hard DriveOSMonitorPower
Seagate Barracuda 250GB Windows 7 Professional 64bit Samsung SW2494 24" Antec Earthpower 650W 
Case
Apevia X-Plorer 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
Intel Q6600 2.40GHZ Gigabyte GA-P35-DS3L MSI GTX570 @ stock Corsair 2GB DDR2 800MHz 
Hard DriveOSMonitorPower
Seagate Barracuda 250GB Windows 7 Professional 64bit Samsung SW2494 24" Antec Earthpower 650W 
Case
Apevia X-Plorer 
  hide details  
Reply
post #12 of 21
Quote:
Originally Posted by BigFan View Post
hmm, interesting. If the OP doesn't mind me asking since I had a similar question:
What about my Q6600? I have it at stock atm and decided to re-oc back to 3Ghz to remove any bottlenecks wwith my GTX570(although I was surprised such could occur at 1080p). I keep getting errors in orthos 18min in running 2 instances, one being blend(cores 0 and 1-tests cpu and memory) and stresscpu(cores 2 and 3-testing only cpu).

I can't increase voltage more since my temps at load hit a max of 68C, mostly hovering around 66C or so and to me, that's getting too high for my comfort. I even ran 3DMark 11 after and before ocing with the difference being 11 points I remember when doing such a oc made a much bigger difference in 06(not sure about vantage), guessing my Q6600 is getting old :/

So, curious what do you guys think?

P.S. OP, if you rather I remove this post, inform me and I'll make another thread
P.S.x2 True increasing IQ and resolution cause a decrease in fps but how is that removing a cpu bottleneck?
Overclocking the CPU helps a lot on 06 because it's old and very easy for a GPU to run, thus it runs at a very high FPS (vs. what it was designed for), and thus the CPU acts as limiting factor to performance even on the GPU tests (what we call CPU bottlenecking). It doesn't help that much on 3dMark11 because it's new and very difficult to run, thus it runs at a much lower FPS, thus the load on the CPU is much lower.

To answer your last PS question, as I assume that's directed at me:

What I'm explaining here is that you'll see people on OCN constantly saying that running at high res (and/or high IQ) will help remove the CPU bottleneck ... which is true ... but the actual mechanism at play is that by raising res, you increase the workload on GPU, thus you lower the FPS, thus you decrease the amount of work per unit of time the CPU has to do in order to keep up, thus the likelihood that the CPU will act as the limiting factor to performance decreases, which in turn is the same thing as saying the CPU bottleneck is being removed
    
CPUMotherboardGraphicsRAM
xeon X5675 6-core @ 4.1ghz (1.29v, 20x205 +ht ) rampage iii extreme msi rx470 gaming X (the $159 budget king) 3 x 2gb corsair xms3 pc12800 (9-9-9-24-1T@1600MHz) 
Hard DriveOptical DriveCoolingOS
hynix 250gb ssd (boot), 2tb deskstar (apps),1tb... plextor px-712sa - still the best optical drive... corsair h8o v2 aio W10 home 
MonitorPowerCaseAudio
asus vw266h 25.5" (1920x1200) abs sl (enermax revolution) * single 70A rail 850w silverstone rv-03 XFi Titanium 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
xeon X5675 6-core @ 4.1ghz (1.29v, 20x205 +ht ) rampage iii extreme msi rx470 gaming X (the $159 budget king) 3 x 2gb corsair xms3 pc12800 (9-9-9-24-1T@1600MHz) 
Hard DriveOptical DriveCoolingOS
hynix 250gb ssd (boot), 2tb deskstar (apps),1tb... plextor px-712sa - still the best optical drive... corsair h8o v2 aio W10 home 
MonitorPowerCaseAudio
asus vw266h 25.5" (1920x1200) abs sl (enermax revolution) * single 70A rail 850w silverstone rv-03 XFi Titanium 
  hide details  
Reply
post #13 of 21
Quote:
Originally Posted by brettjv View Post
Overclocking the CPU helps a lot on 06 because it's old and very easy for a GPU to run, thus it runs at a very high FPS (vs. what it was designed for), and thus the CPU acts as limiting factor to performance even on the GPU tests (what we call CPU bottlenecking). It doesn't help that much on 3dMark11 because it's new and very difficult to run, thus it runs at a much lower FPS, thus the load on the CPU is much lower.

To answer your last PS question, as I assume that's directed at me:

What I'm explaining here is that you'll see people on OCN constantly saying that running at high res (and/or high IQ) will help remove the CPU bottleneck ... which is true ... but the actual mechanism at play is that by raising res, you increase the workload on GPU, thus you lower the FPS, thus you decrease the amount of work per unit of time the CPU has to do in order to keep up, thus the likelihood that the CPU will act as the limiting factor to performance decreases, which in turn is the same thing as saying the CPU bottleneck is being removed
hmm, that's what I was thinking, thought you meant that the ratio of gpu:cpu is still 1:1 or whatever, not that gpu does more work, nevermind

As for 06, I see your point. Thanks for the info

Little update: 48min and no errors with the cpu at 100% load. Likely memory error, but, my memory is working fine at 800mhz, how come 803mhz is causing errors? Used to be fine at 833mhz as well when I oc it, guess its age again

Edit2: One of the instances froze, so, trying a mem test now :/
Edited by BigFan - 1/4/11 at 3:23pm
    
CPUMotherboardGraphicsRAM
Intel Q6600 2.40GHZ Gigabyte GA-P35-DS3L MSI GTX570 @ stock Corsair 2GB DDR2 800MHz 
Hard DriveOSMonitorPower
Seagate Barracuda 250GB Windows 7 Professional 64bit Samsung SW2494 24" Antec Earthpower 650W 
Case
Apevia X-Plorer 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
Intel Q6600 2.40GHZ Gigabyte GA-P35-DS3L MSI GTX570 @ stock Corsair 2GB DDR2 800MHz 
Hard DriveOSMonitorPower
Seagate Barracuda 250GB Windows 7 Professional 64bit Samsung SW2494 24" Antec Earthpower 650W 
Case
Apevia X-Plorer 
  hide details  
Reply
post #14 of 21
Quote:
Originally Posted by BigFan View Post
hmm, that's what I was thinking, thought you meant that the ratio of gpu:cpu is still 1:1 or whatever, not that gpu does more work, nevermind
Yeah, what I'm saying is the increasing the IQ/res increases the absolute workload on the GPU, but does not change the relative workload between the CPU and GPU. That ratio basically stays the same regardless of the resolution (and is properly understood as describing the level of CPU dependency, or the amount of work required of the CPU per frame rendered, of the particular application in question ... of course, even within a given application, this ratio is in a constant state of flux).

Ergo (that's my word of the day), the only mechanism by which one reduces the probability of the CPU acting as a limiting factor to performance is by reducing the frequency with which the GPU demands data from the CPU, which is determined by the FPS that the GPU is running at.

Well, that's not always 100% true because there are sometimes settings in games with which one can actually lower the CPU dependency, such as turning down the physics quality, which is something that runs on the CPU (unless it's GPU-PhysX, of course).

But bottom-line, if increasing res/IQ did NOT lower the FPS, it would also NOT act to reduce CPU bottlenecking ... because the level of CPU dependency (amount of CPU work required per frame rendered) is not going to change either. And if it was going to change in either direction, it would be upwards, not downwards.
Edited by brettjv - 1/4/11 at 3:34pm
    
CPUMotherboardGraphicsRAM
xeon X5675 6-core @ 4.1ghz (1.29v, 20x205 +ht ) rampage iii extreme msi rx470 gaming X (the $159 budget king) 3 x 2gb corsair xms3 pc12800 (9-9-9-24-1T@1600MHz) 
Hard DriveOptical DriveCoolingOS
hynix 250gb ssd (boot), 2tb deskstar (apps),1tb... plextor px-712sa - still the best optical drive... corsair h8o v2 aio W10 home 
MonitorPowerCaseAudio
asus vw266h 25.5" (1920x1200) abs sl (enermax revolution) * single 70A rail 850w silverstone rv-03 XFi Titanium 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
xeon X5675 6-core @ 4.1ghz (1.29v, 20x205 +ht ) rampage iii extreme msi rx470 gaming X (the $159 budget king) 3 x 2gb corsair xms3 pc12800 (9-9-9-24-1T@1600MHz) 
Hard DriveOptical DriveCoolingOS
hynix 250gb ssd (boot), 2tb deskstar (apps),1tb... plextor px-712sa - still the best optical drive... corsair h8o v2 aio W10 home 
MonitorPowerCaseAudio
asus vw266h 25.5" (1920x1200) abs sl (enermax revolution) * single 70A rail 850w silverstone rv-03 XFi Titanium 
  hide details  
Reply
post #15 of 21
Quote:
Originally Posted by brettjv View Post
Yeah, what I'm saying is the increasing the IQ/res increases the absolute workload on the GPU, but does not change the relative workload between the CPU and GPU. That ratio basically stays the same regardless of the resolution (and is properly understood as describing the level of CPU dependency, or the amount of work required of the CPU per frame rendered) .

Ergo (that's my word of the day), the only mechanism by which one reduces the probability of the CPU acting as a limiting factor to performance is by reducing the frequency with which the GPU demands data from the CPU, which is determined by the FPS that the GPU is running at.

Well, that's not always 100% true because there are sometimes settings in games with which one can actually lower the CPU dependency, such as turning down the physics quality, which is something that runs on the CPU (unless it's GPU-PhysX, of course).
ok, makes sense now, thanks for the info. +REP for that big writeup

Running a mix of blend and stresscpu, hoping to pass 18min now to see if I pass the error time. Any advice on the temps or how bad of a bottleneck my Q6600 is?

Edit: Well, there it goes XD Error(Vcoul_ref=2.574246, Vcoul=2.574246) at 9 mins and 8 second :/
Edit2: Makes sense^^
Edited by BigFan - 1/4/11 at 3:36pm
    
CPUMotherboardGraphicsRAM
Intel Q6600 2.40GHZ Gigabyte GA-P35-DS3L MSI GTX570 @ stock Corsair 2GB DDR2 800MHz 
Hard DriveOSMonitorPower
Seagate Barracuda 250GB Windows 7 Professional 64bit Samsung SW2494 24" Antec Earthpower 650W 
Case
Apevia X-Plorer 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
Intel Q6600 2.40GHZ Gigabyte GA-P35-DS3L MSI GTX570 @ stock Corsair 2GB DDR2 800MHz 
Hard DriveOSMonitorPower
Seagate Barracuda 250GB Windows 7 Professional 64bit Samsung SW2494 24" Antec Earthpower 650W 
Case
Apevia X-Plorer 
  hide details  
Reply
post #16 of 21
You know I really wish benchmarks done by reviews sites included more on minimum/average fps.
Main rig
(15 items)
 
  
CPUMotherboardGraphicsRAM
6700k 4.6ghz GIGABYTE GA-Z170X-UD3 Asus GTX 1070 STRIX 16GB G.skill ripjaw DDR4 
Hard DriveCoolingOSMonitor
2x 520 256gb SSD / 1x 256gb muskin SSD/ 1x WD 2... NH-DS15s Win10 27" 1440p 100hz, IPS 23' 
KeyboardPowerMouseAudio
Quickfire Stealth 850tx Rival 300 hd600 
  hide details  
Reply
Main rig
(15 items)
 
  
CPUMotherboardGraphicsRAM
6700k 4.6ghz GIGABYTE GA-Z170X-UD3 Asus GTX 1070 STRIX 16GB G.skill ripjaw DDR4 
Hard DriveCoolingOSMonitor
2x 520 256gb SSD / 1x 256gb muskin SSD/ 1x WD 2... NH-DS15s Win10 27" 1440p 100hz, IPS 23' 
KeyboardPowerMouseAudio
Quickfire Stealth 850tx Rival 300 hd600 
  hide details  
Reply
post #17 of 21
The proper way to determine the degree to which the CPU is acting as a bottleneck is to run a benchmark at your current clocks, then OC the proc and run the benchmark again. Observe the % change in FPS that resulted from the CPU overclock, and divide this by the % change in clock speed.

If you observed a 10% FPS increase from a 10% increase in CPU clocks, then you have a large (as in, probably all the frames rendered were being CPU-limited, and/or all those that were cpu-limited were severely so ... keep in mind we're dealing with averages here) bottleneck.

If you got a 1% increase in FPS from a 10% increase in clocks, then you have a very small (infrequent, and/or not severe in terms of all the frames that were rendered during the test) CPU bottleneck.

Note that whatever the level of BN, it describes the situation at the lower of the two CPU clocks.

But as has been observed upthread, overclocking the q6600 does mostly nothing afa the GPU score on 3dMark11, but does a lot to affect the GPU score on 3dMark06. The same sort of variances will apply to everything you ever run.

Ergo (there's that word again ) whatever you observe will be be entirely specific to the particular test at hand, i.e. what application (bench/game) were you running (which affects both the level of CPU dependency, and the absolute workload on the GPU), and what settings were you running it at (which only affects the absolute workload on the GPU).

Does this all make sense?
Edited by brettjv - 1/4/11 at 4:09pm
    
CPUMotherboardGraphicsRAM
xeon X5675 6-core @ 4.1ghz (1.29v, 20x205 +ht ) rampage iii extreme msi rx470 gaming X (the $159 budget king) 3 x 2gb corsair xms3 pc12800 (9-9-9-24-1T@1600MHz) 
Hard DriveOptical DriveCoolingOS
hynix 250gb ssd (boot), 2tb deskstar (apps),1tb... plextor px-712sa - still the best optical drive... corsair h8o v2 aio W10 home 
MonitorPowerCaseAudio
asus vw266h 25.5" (1920x1200) abs sl (enermax revolution) * single 70A rail 850w silverstone rv-03 XFi Titanium 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
xeon X5675 6-core @ 4.1ghz (1.29v, 20x205 +ht ) rampage iii extreme msi rx470 gaming X (the $159 budget king) 3 x 2gb corsair xms3 pc12800 (9-9-9-24-1T@1600MHz) 
Hard DriveOptical DriveCoolingOS
hynix 250gb ssd (boot), 2tb deskstar (apps),1tb... plextor px-712sa - still the best optical drive... corsair h8o v2 aio W10 home 
MonitorPowerCaseAudio
asus vw266h 25.5" (1920x1200) abs sl (enermax revolution) * single 70A rail 850w silverstone rv-03 XFi Titanium 
  hide details  
Reply
post #18 of 21
Quote:
Originally Posted by brettjv View Post
The proper way to determine the degree to which the CPU is acting as a bottleneck is to run a benchmark at your current clocks, then OC the proc and run the benchmark again. Observe the % change in FPS that resulted from the CPU overclock, and divide this by the % change in clock speed.

If you observed a 10% FPS increase from a 10% increase in CPU clocks, then you have a large (as in, probably all the frames rendered were being CPU-limited, and/or all those that were cpu-limited were severely so ... keep in mind we're dealing with averages here) bottleneck.

If you got a 1% increase in FPS from a 10% increase in clocks, then you have a very small (infrequent, and/or not severe in terms of all the frames that were rendered during the test) CPU bottleneck.


Does this all make sense?
Of course, it makes sense.

Let me provide a graphical way for the above bolded part:

Plot a graph with CPU frequency as x-axis and framerate as y-axis.

If the slope of the curve at a particular frequency is steep, then the bottleneck is large at that frequency.

If the slope is small (ie leveling off), then the bottleneck is small at that frequency.
Edited by windfire - 1/4/11 at 4:40pm
post #19 of 21
Quote:
Originally Posted by windfire View Post
Of course, it makes sense.

Let me provide a graphical way for the above bolded part:

Plot a graph with CPU frequency as x-axis and framerate as y-axis.

If the slope of the curve at a particular frequency is steep, then the bottleneck is large at that frequency.

If the slope is small (ie leveling off), then the bottleneck is small at that frequency.
Of course it makes sense to YOU, Windy. You always know what I'm talking about

For an illustration of the sort of graphing that Windfire is referring to, see this article (which also has a good explanation of bottleneck on the first page):

http://benchmarkextreme.com/Articles...alysis/P5.html

The following article posted on the same site is also very good afa the subject of BN's goes, and what a dynamic situation it is depending on the test and settings being used:

http://benchmarkextreme.com/Articles...alysis/P1.html

If you read through both these articles, you'll know a heck of a lot more about bottlenecks than you did before, I can promise ya that
Edited by brettjv - 1/4/11 at 5:20pm
    
CPUMotherboardGraphicsRAM
xeon X5675 6-core @ 4.1ghz (1.29v, 20x205 +ht ) rampage iii extreme msi rx470 gaming X (the $159 budget king) 3 x 2gb corsair xms3 pc12800 (9-9-9-24-1T@1600MHz) 
Hard DriveOptical DriveCoolingOS
hynix 250gb ssd (boot), 2tb deskstar (apps),1tb... plextor px-712sa - still the best optical drive... corsair h8o v2 aio W10 home 
MonitorPowerCaseAudio
asus vw266h 25.5" (1920x1200) abs sl (enermax revolution) * single 70A rail 850w silverstone rv-03 XFi Titanium 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
xeon X5675 6-core @ 4.1ghz (1.29v, 20x205 +ht ) rampage iii extreme msi rx470 gaming X (the $159 budget king) 3 x 2gb corsair xms3 pc12800 (9-9-9-24-1T@1600MHz) 
Hard DriveOptical DriveCoolingOS
hynix 250gb ssd (boot), 2tb deskstar (apps),1tb... plextor px-712sa - still the best optical drive... corsair h8o v2 aio W10 home 
MonitorPowerCaseAudio
asus vw266h 25.5" (1920x1200) abs sl (enermax revolution) * single 70A rail 850w silverstone rv-03 XFi Titanium 
  hide details  
Reply
post #20 of 21
Thread Starter 
Wow, great additional info brettjv. Guess I didnt know everything worth knowing about bottlenecks after all Oh, and good idea windfire. Guess I can put that Excel at home to use after all. Hihi.
    
CPUMotherboardGraphicsRAM
i7 4790k @ 4.5ghz MSI Z97 Gaming 7 Gigabyte Aorus 1080 ti@2.012ghz 16 gigs of Corsair Vengeance@1600 
Hard DriveCoolingOSMonitor
1xSamsung SSD 500 mb, 1xCorsair M550 SSD 1 tb, ... Cryorig R1 Universal Windows 8.1 Pro 64-bit 27" Asus pb278q 
KeyboardPowerCaseMouse
Logitech G510 Corsair TX750W Corsair Obsidian 550d Logitech gx9 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
i7 4790k @ 4.5ghz MSI Z97 Gaming 7 Gigabyte Aorus 1080 ti@2.012ghz 16 gigs of Corsair Vengeance@1600 
Hard DriveCoolingOSMonitor
1xSamsung SSD 500 mb, 1xCorsair M550 SSD 1 tb, ... Cryorig R1 Universal Windows 8.1 Pro 64-bit 27" Asus pb278q 
KeyboardPowerCaseMouse
Logitech G510 Corsair TX750W Corsair Obsidian 550d Logitech gx9 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: NVIDIA
Overclock.net › Forums › Graphics Cards › NVIDIA › Any point of overclocking a GTX570 with only 3.2 ghz q9550?