GIGABYTE GTX 9xx H2O/AIR BIOS Tweaking ?(ô?ô)? - Page 407 - Overclock.net - An Overclocking Community

Forum Jump: 

GIGABYTE GTX 9xx H2O/AIR BIOS Tweaking ?(ô?ô)?

Reply
 
Thread Tools
post #4061 of 7850 (permalink) Old 03-20-2016, 05:20 AM
New to Overclock.net
 
Edkiefer's Avatar
 
Join Date: Feb 2013
Posts: 1,441
Rep: 115 (Unique: 62)
Quote:
Originally Posted by Olorin View Post

Well, tried to get started on it a little bit tonight, but running into some weird error.  First, GPU-Z would stop responding when I tried to Save my existing Bios to file.  A quick Google revealed that you're supposed to disable the card in device manager first.  OK, no freezing GPU-Z, but then I see this error:




I guess it's time to step back and read about what the heck I'm doing before I get ahead of myself!


Searched a little in this thread, then went back to this thread:
https://www.overclock.net/t/1521334/official-nvflash-with-certificate-checks-bypassed-for-gtx-950-960-970-980-980ti-titan-x/0_20 

Which of course led to where I need to be:
https://www.overclock.net/t/1523391/easy-nvflash-guide-with-pictures-for-gtx-970-980/0_20

Guess I'll dig in and have some fun after I actually figure out what the heck I'm doing!  G'nite, y'all!!! sleepysmileyanim.gif
Make sure your using latest GPU-Z, it should disable card before extracting the bios to save. In older versions you had to manually disable card in device manager, just like you do before flashing card (there no need to uninstall drivers).

Edkiefer is offline  
Sponsored Links
Advertisement
 
post #4062 of 7850 (permalink) Old 03-20-2016, 07:49 AM
New to Overclock.net
 
0xzz's Avatar
 
Join Date: Aug 2015
Location: Germany
Posts: 267
Rep: 15 (Unique: 6)
Quote:
Originally Posted by Olorin View Post

Well, tried to get started on it a little bit tonight, but running into some weird error.  First, GPU-Z would stop responding when I tried to Save my existing Bios to file.  A quick Google revealed that you're supposed to disable the card in device manager first.  OK, no freezing GPU-Z, but then I see this error:




I guess it's time to step back and read about what the heck I'm doing before I get ahead of myself!


Searched a little in this thread, then went back to this thread:
https://www.overclock.net/t/1521334/official-nvflash-with-certificate-checks-bypassed-for-gtx-950-960-970-980-980ti-titan-x/0_20 

Which of course led to where I need to be:
https://www.overclock.net/t/1523391/easy-nvflash-guide-with-pictures-for-gtx-970-980/0_20

Guess I'll dig in and have some fun after I actually figure out what the heck I'm doing!  G'nite, y'all!!! sleepysmileyanim.gif

make sure you are running it as admin and no anti-virus program is interfering

0xzz is offline  
post #4063 of 7850 (permalink) Old 03-20-2016, 08:47 AM
New to Overclock.net
 
DeathAngel74's Avatar
 
Join Date: Aug 2014
Location: Next to the evil monkey in your closet!
Posts: 5,380
Rep: 267 (Unique: 141)
@0xzz
How do you like the 6700k/maximus viii hero combo? I just ordered mine along with 32gb DDR 3200MHz. Wondering if you delidded? Pros/cons in general. Thanks
DeathAngel74 is offline  
Sponsored Links
Advertisement
 
post #4064 of 7850 (permalink) Old 03-20-2016, 08:51 AM
New to Overclock.net
 
0xzz's Avatar
 
Join Date: Aug 2015
Location: Germany
Posts: 267
Rep: 15 (Unique: 6)
Quote:
Originally Posted by DeathAngel74 View Post

@0xzz
How do you like the 6700k/maximus viii hero combo? I just ordered mine along with 32gb DDR 3200MHz. Wondering if you delidded? Pros/cons in general. Thanks

i like it a lot, had no problems so far but i think i got a crappy cpu, gets really hot and im not able to push it to 4.8ghz. im thinking about delidding but im reading about people who delidded and they cracked their DIE and killed their chips after putting them back together so i dont know if i will do it

0xzz is offline  
post #4065 of 7850 (permalink) Old 03-20-2016, 08:54 AM
New to Overclock.net
 
DeathAngel74's Avatar
 
Join Date: Aug 2014
Location: Next to the evil monkey in your closet!
Posts: 5,380
Rep: 267 (Unique: 141)
Get a RocketCool delidding tool.
$34.95

No vice, hammer, or razors needed.
DeathAngel74 is offline  
post #4066 of 7850 (permalink) Old 03-20-2016, 09:06 AM - Thread Starter
Maximum_Unleashed
 
Laithan's Avatar
 
Join Date: Mar 2015
Location: United States
Posts: 3,699
Rep: 500 (Unique: 265)
Quote:
Originally Posted by Olorin View Post

Well, finally got around to testing for game stability on my 1542 core, 8000 mem overclock and disappointing, Fallout 4 crashed in just a few minutes.  So I dropped to 1532 core and 7806 memory, and it played well for about 30 minutes before crashing. Both crashes gave the Nvidia driver stopped working and has restarted message from Windows 10.  gunner.gif

Still, this is all with stock bios.  I definitely plan to flash to the custom bios for 970 G1.  I did do the testing with the case closed, but only after adding 2 more fans. Now I have two 120mm fans exhausting upwards, one exhausting out the back, plus the CPU cooler, and 5 fans blowing air in.  2 in the front grills, 2 behind the HDD cages, 1 on the side blowing towards the PSU and GPU.  I don't think temperature is the problem, as Tmax was 64C.  I think it's voltage/power.  A short Kombustor test showed a Power perfcap. Heaven and gaming show Vrel and Vop perfcaps, but not power.  I ordered the Everflow 80mm fan with LED display for the ghetto cooling option. It may not fit, though because the Thermalright Macho Rev B is so freakin massive.  I think I measured 90 or 95mm room from the edge of the Macho's fan to the end of the backplate on the 970GTX. If it's too big, I'll save it for one of my other rigs and order a 60mm fan.

Tomorrow I'll read up on how to flash the bios and see how things go with the Vrel, Vops perfcaps removed... looking forward to it.  Just too tired tonight!  Maybe someday I'll finish reading every post in this thread, but it's a bit lengthy! But lots of cool info. Thanks for all the hard work, Laithan, and everyone else who's contributed.  thumbsupsmiley.png

BTW, Laithan, your case and watercool setup looks sweet!  Out of curiosity, do you have to disassemble the 40mm fans to dye them?  They look sweet dyed to match the build.


Thank you sir smile.gif

Good idea, take a good baseline before flashing so you can have a very clear and detailed understanding. As far as you crashing that's called a TDR. It's a recovery mechanism supported in the O/S as well as the driver. When the GPU can't handle things anymore it basically quits..and instead of just locking up like the old days, now we are kicked out and sent back to the desktop. When this happens even if things appear to be OK, you MUST reboot to make things right again. Trust me on this one smile.gif. It is quite normal to crash whether a stock or mod BIOS... you will crash a LOT while learning your GPU. smile.gif

Some advice for you and others: What really needs to happen is to maximize your overclocking potential physically and then to learn the specific limits of your GPU virtually in software. This has a lot to do with managing temps/airflow and isolating memory and GPU overclocking limits individually and then combined.

Stability testing is quite easy to get WRONG actually. I'm not saying you did or did not perform correct stability testing (I speak to the audience a lot) I am just saying that it is important to REALLY understand how GPU workload fluctuates from one app/game to another. Quite frankly our GPUs have different limits based on the application's ability to push the GPU. This means that for a less demanding graphical workload I am able to clock my GPU higher. A more complex (next-gen) GPU workload would crash at the same settings as the less demanding one. You can get the idea that we can "chase our tail" a lot when stability testing if this simple concept is not understood because if we have absolutely zero problems in one app/game this does NOT prove or disprove that same level of stability in other app/games, unless the app/game that is chosen fully utilizes the GPU at full load for voltage, power and an extremely complex and demanding GPU load PLUS pushing all of your CPU cores hard ALL AT THE SAME TIME... That's really the true test for stability.

So you have to have the "right tools for the job" smile.gif. What most people do is just run a benchmark or play a game.. This is obviously going to be the way that we test the GPU but there needs to be an understanding beyond that if we want true stability testing to occur. Many of the app/games will push the GPU hard but not the CPU.. or might push the GPU and CPU hard but isn't using very much GPU power (the power used is based on the GPU workload, not utilization). You can calculate and estimate the POTENTIAL power draw based on the total inventory of your system components. Furmark/Kombuster is designed to test ONLY GPU power draw (maximum possible) which is good to know however it's simply VERY far from real world power usage in ANY app/game. Unless you have a wall meter you will not realize this. CPU workload is also not factored. For this reason it isn't of much value for stability testing because nothing you will ever do will ever require as much power as this synthetic test..not even close.. It's like testing your car for stability at 200Mph/Kph when you are never going to drive over 110Mph/Kph.

There are 4 major factors that play a role when finding your GPU's highest stability:
(1) You need to be SIMULTANEOUSLY pushing your CPU, your system memory and in addition your GPUs need to not just have high utilization but they also need to be drawing a lot of power, running at max voltage and have a complex GPU workload to truly test stability. App/Games that do this are few and far between!

(2) Throttling - There are three types. Insufficient voltage, insufficient power and overheating can all lead to throttling. A STOCK BIOS will run into voltage and power throttling no matter what you do. You need a MOD BIOS to open your GPU to it's full potential and avoid your GPU "jumping around" and constantly throttling. Watch your temps!!

(3) Power Supply - Not all app/games are going to draw the maximum power that your GPU is capable of pulling. Do yourself a favor, spend the MOST money on your POWER SUPPLY when building/buying a PC! This IS the MOST important component in your system. If power is ever unreliable and/or fluctuates expect to have a lot of inconsistent problems that will drive you crazy and will be very difficult to troubleshoot. Pay attention to the PSU requirements especially with a MOD BIOS.

(4) You have 2 things to overclock.. Your GPU and your GPU MEMORY. Do NOT try to find your maximum limit on BOTH at the same time. Leave your memory on STOCK and overclock your GPU only at first.. Test for a lengthy period.. one test is not enough.


My recommendation:
(1) You MUST be testing @ 4K otherwise you aren't pushing your GPU enough (DSR 4x if you only have a 1080p monitor)
(2) Use FFXIV free benchmark. Set the preset to MAXIMUM and change the resolution to 4K. This test is a PERFECT example of a test that will SIMULTANEOUSLY draw a LOT of power, will use max voltage, will push your temps high and give your CPU a nice workout too tongue.gif


GTG will ramble more later smile.gif

GIGABYTE GTX 9xx G1 GAMING BIOS Tweaking
  ̿̿ ̿̿ ̿̿ ̿̿̿'̿'\̵͇̿̿\з=༼ຈل͜ຈ༽=ε/̵͇̿̿/'̿'̿ ̿ ̿̿ ̿̿ ̿̿  
░▒▓│ FORKAY‼ «  » WΘΘT ‼ │▓▒░
Xeon E5-1680 V2 (IVY-E) Inside


Laithan is offline  
post #4067 of 7850 (permalink) Old 03-20-2016, 10:16 AM
New to Overclock.net
 
Olorin's Avatar
 
Join Date: Sep 2006
Location: Aztec, NM
Posts: 120
Rep: 4 (Unique: 3)
Quote:
Originally Posted by Laithan View Post


Thank you sir smile.gif

Good idea, take a good baseline before flashing so you can have a very clear and detailed understanding. As far as you crashing that's called a TDR. It's a recovery mechanism supported in the O/S as well as the driver. When the GPU can't handle things anymore it basically quits..and instead of just locking up like the old days, now we are kicked out and sent back to the desktop. When this happens even if things appear to be OK, you MUST reboot to make things right again. Trust me on this one smile.gif. It is quite normal to crash whether a stock or mod BIOS... you will crash a LOT while learning your GPU. smile.gif

Some advice for you and others: What really needs to happen is to maximize your overclocking potential physically and then to learn the specific limits of your GPU virtually in software. This has a lot to do with managing temps/airflow and isolating memory and GPU overclocking limits individually and then combined.

Stability testing is quite easy to get WRONG actually. I'm not saying you did or did not perform correct stability testing (I speak to the audience a lot) I am just saying that it is important to REALLY understand how GPU workload fluctuates from one app/game to another. Quite frankly our GPUs have different limits based on the application's ability to push the GPU. This means that for a less demanding graphical workload I am able to clock my GPU higher. A more complex (next-gen) GPU workload would crash at the same settings as the less demanding one. You can get the idea that we can "chase our tail" a lot when stability testing if this simple concept is not understood because if we have absolutely zero problems in one app/game this does NOT prove or disprove that same level of stability in other app/games, unless the app/game that is chosen fully utilizes the GPU at full load for voltage, power and an extremely complex and demanding GPU load PLUS pushing all of your CPU cores hard ALL AT THE SAME TIME... That's really the true test for stability.

So you have to have the "right tools for the job" smile.gif. What most people do is just run a benchmark or play a game.. This is obviously going to be the way that we test the GPU but there needs to be an understanding beyond that if we want true stability testing to occur. Many of the app/games will push the GPU hard but not the CPU.. or might push the GPU and CPU hard but isn't using very much GPU power (the power used is based on the GPU workload, not utilization). You can calculate and estimate the POTENTIAL power draw based on the total inventory of your system components. Furmark/Kombuster is designed to test ONLY GPU power draw (maximum possible) which is good to know however it's simply VERY far from real world power usage in ANY app/game. Unless you have a wall meter you will not realize this. CPU workload is also not factored. For this reason it isn't of much value for stability testing because nothing you will ever do will ever require as much power as this synthetic test..not even close.. It's like testing your car for stability at 200Mph/Kph when you are never going to drive over 110Mph/Kph.

There are 4 major factors that play a role when finding your GPU's highest stability:
(1) You need to be SIMULTANEOUSLY pushing your CPU, your system memory and in addition your GPUs need to not just have high utilization but they also need to be drawing a lot of power, running at max voltage and have a complex GPU workload to truly test stability. App/Games that do this are few and far between!

(2) Throttling - There are three types. Insufficient voltage, insufficient power and overheating can all lead to throttling. A STOCK BIOS will run into voltage and power throttling no matter what you do. You need a MOD BIOS to open your GPU to it's full potential and avoid your GPU "jumping around" and constantly throttling. Watch your temps!!

(3) Power Supply - Not all app/games are going to draw the maximum power that your GPU is capable of pulling. Do yourself a favor, spend the MOST money on your POWER SUPPLY when building/buying a PC! This IS the MOST important component in your system. If power is ever unreliable and/or fluctuates expect to have a lot of inconsistent problems that will drive you crazy and will be very difficult to troubleshoot. Pay attention to the PSU requirements especially with a MOD BIOS.

(4) You have 2 things to overclock.. Your GPU and your GPU MEMORY. Do NOT try to find your maximum limit on BOTH at the same time. Leave your memory on STOCK and overclock your GPU only at first.. Test for a lengthy period.. one test is not enough.


My recommendation:
(1) You MUST be testing @ 4K otherwise you aren't pushing your GPU enough (DSR 4x if you only have a 1080p monitor)
(2) Use FFXIV free benchmark. Set the preset to MAXIMUM and change the resolution to 4K. This test is a PERFECT example of a test that will SIMULTANEOUSLY draw a LOT of power, will use max voltage, will push your temps high and give your CPU a nice workout too tongue.gif


GTG will ramble more later smile.gif

 

Laithan,

 

Thanks so much for the very, very detailed reply.  I have downloaded the FFXIV benchmark and tried it once, but not as DSR 4x, which I'll try sometime after I carefully study how to flash the mod bios.  

 

And you're absolutely right that I took shortcuts :blushsmil and overclocked the core and memory together instead of independently.  I haven't even found the limits of the memory overclock, I just stopped when I got to +500.  I think I was OC'ing memory only at that point, but curiosity/impatience got the better of me and I decided to try my core overclock with the +500 memory I had reached up to that point.  I was thrilled that it appeared stable in Heaven at +150 core and +500 memory.  I should have returned to +0 core and continued finding the memory limit alone.  I think it was probably past midnight, though and I picked up again the next day, but started again with the "profile" I'd created of +150 core and +500 memory.

 

So, going forward I'll:

  1. Flash the mod bios (after studying the steps, backing up my bios)
  2. Test my core only with mod bios (using FFXIV at DSR 4x instead of simple passes of Unigine Heaven, I guess?)
  3. Test my memory only with mod bios
  4. Test core and memory together with FFXIV at DSR 4x
  5. Then start game stability testing

 

I'll let the thread know how things go along the way! :kungfu:


GIGABYTE GTX 9xx/10xx G1 GAMING H2O/AIR BIOS Tweaking ┌(ô益ô)┐

 

Metro DataVac Club

 

Core i5 4690K @ 4.4GHz, 1.32V Vcore, 2.172V Vrin

Gigabyte G1 Gaming GTX 970, 1544Mhz core, 8010MHz memory

Olorin is offline  
post #4068 of 7850 (permalink) Old 03-20-2016, 10:30 AM
New to Overclock.net
 
0xzz's Avatar
 
Join Date: Aug 2015
Location: Germany
Posts: 267
Rep: 15 (Unique: 6)
Quote:
Originally Posted by DeathAngel74 View Post

Get a RocketCool delidding tool.
$34.95

No vice, hammer, or razors needed.

no the damage was not done while delidding, the DIE cracked after putting pressure on the IHS after placing it back on the pcb

0xzz is offline  
post #4069 of 7850 (permalink) Old 03-20-2016, 10:31 AM
New to Overclock.net
 
Olorin's Avatar
 
Join Date: Sep 2006
Location: Aztec, NM
Posts: 120
Rep: 4 (Unique: 3)
Quote:
Originally Posted by 0xzz View Post


make sure you are running it as admin and no anti-virus program is interfering

You hit it on the head.  I hadn't gone into the compatibility settings and enabled "Run as Administrator".  Once I did that, it worked just fine.  No need to disable the display adapter with Device Manager, just like you said.  Thanks for clearing that up for me!

 

Edkiefer,

Thanks for the advice as well. I am running the latest 0.8.7 ROG version (just for the looks). I just wasn't running it "right".

 

:doh:


GIGABYTE GTX 9xx/10xx G1 GAMING H2O/AIR BIOS Tweaking ┌(ô益ô)┐

 

Metro DataVac Club

 

Core i5 4690K @ 4.4GHz, 1.32V Vcore, 2.172V Vrin

Gigabyte G1 Gaming GTX 970, 1544Mhz core, 8010MHz memory

Olorin is offline  
post #4070 of 7850 (permalink) Old 03-20-2016, 11:36 AM
New to Overclock.net
 
pepi93's Avatar
 
Join Date: Aug 2013
Posts: 196
Rep: 0
[/quote]
Quote:
Originally Posted by Laithan View Post


My recommendation:
(1) You MUST be testing @ 4K otherwise you aren't pushing your GPU enough (DSR 4x if you only have a 1080p monitor)
(2) Use FFXIV free benchmark. Set the preset to MAXIMUM and change the resolution to 4K. This test is a PERFECT example of a test that will SIMULTANEOUSLY draw a LOT of power, will use max voltage, will push your temps high and give your CPU a nice workout too tongue.gif

Great post!

I ran the benchmark as suggested but I did notice that the stress on the CPU wasn't that much? Usually in the 30's%, sometimes it would skip to 40's % and even mid 50's% but rarely. The stress on the GPU was quite good though, several times drawing full power.

Did I do something wrong in my setup that it didn't push my CPU?

Phanteks Enthoo Pro M SE, EVGA 1000W Platinum PSU
Phanteks Fans, RGB Halos, RGB strips
Asus Maximus VIII Hero, i7-6700k, Corsair H100iV2
Corsair Vengeance 2x8gb, LPX 3000mhz
Gigabyte 980ti G1, 850 EVO 250GB, OCZ Agility 120gb, 840 EVO 500GB, Seagate Barracuda 3TB HDD
BenQ BL3200PT, BenQ 24 RL2460HT
Klipsch Promedia 2.1 THX Speakers, AKG-K550 Reference Headphones, Fiio E11 amp
Roccat [Kone XTD Mouse & Ryos MK Pro Keyboard]
pepi93 is offline  
Reply

Quick Reply
Message:
Options

Register Now

In order to be able to post messages on the Overclock.net - An Overclocking Community forums, you must first register.
Please enter your desired user name, your email address and other required details in the form below.
User Name:
If you do not want to register, fill this field only and the name will be used as user name for your post.
Password
Please enter a password for your user account. Note that passwords are case-sensitive.
Password:
Confirm Password:
Email Address
Please enter a valid email address for yourself.
Email Address:

Log-in



Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)
 
Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may post new threads
You may post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off