Go Back   Hardware Canucks > PC BUILDERS & TWEAKERS CORNER > Overclocking, Tweaking and Benchmarking

    
Reply
 
LinkBack Thread Tools Display Modes
  #61 (permalink)  
Old August 27, 2015, 03:27 AM
trodas's Avatar
MVP
 
Join Date: Feb 2008
Location: Czech republic
Posts: 394

My System Specs

Default

Well, hopefully some people join at stages, where I cannot help at all. I can provide - hopefully valid - slowest GPUPI run: ATM I got 324h on the Batch 18, so only two more loops (~48h) and we know the result... No screen clear now, all results are bellow visible:



Since the 324h mark was hit near end of 26.8., then in about 22 to 23h today I shall be getting Batch 19, so one way or another, this is going to end.


Also I acquired back the Asus TXP4-X mainboard I recapped for friend in 2008, Socket 7, with rare bug - possibility to enable 7.14MHz FSB, resulting in 10.7MHz for CPU at 1.5x multiplier:

CPU-Z VALIDATOR

That should take care the lowest CPU clocks for Intel (P90) and AMD (K5) :D If my experiment with the Pentium Overdrive (if I can acquire it!) works, then Intel could get at low, as 7.14MHz CPU clocks.

The slowest Aquamark test is also easy, if the validation will go on, then I bet no-one can come even close to me, because it is running since 13.8. and I got only 2778 frames, as of now. So hardly can anyone come even close to that...

The rest are on YOU, guys. I lack the necessary HW. (ATM I still searching for PCI S3 Trio64 card to get the Asus TXP4-X to even visibly post, beeps and leds on the keyboard are not enought for me :) ...K5 I have promised, but not delivered (yet) and Pentium Overdrive is not yet ordered from eBay even... )
__________________
"It is dangerous to be right in matters on which the established authorities are wrong." - Voltaire
"I believe that all the people who stand to profit by a war and who help provoke it should be shot on the first day it starts..." - Hemingway :) my config - my caps
Reply With Quote
  #62 (permalink)  
Old August 27, 2015, 08:40 AM
Dead Things's Avatar
Hall Of Fame
F@H
 
Join Date: Oct 2008
Location: Centre of the Universe
Posts: 1,969

My System Specs

Default

Good luck with the TXP4-X trodas! And don't forget to use the latest CPU-Z (1.73) on WinXP for the validation.

And I've got my fingers crossed for no solar flares before you can complete your Aquamark and GPUPI runs!
__________________
Think you can overclock? Then show us what you got!
Join the Hardware Canucks Overclocking team today!

Follow my benching, folding, mining and miscellaneous shenanigans @dt_oc
Reply With Quote
  #63 (permalink)  
Old August 27, 2015, 12:07 PM
trodas's Avatar
MVP
 
Join Date: Feb 2008
Location: Czech republic
Posts: 394

My System Specs

Default

Installing WinXP on the mobo will be painfull at least. IIRC WinXP refused to boot (as well, as Win2k, witch is enought for latest CPU-Z) at 10.7MHz. Some people claim, that when installed on the clock frequency, then it boot, but that would be a lenghty process...!
Windows XP Minimal-Requirement-Test

Such install could easily take a week or so. Even if I strip down the install by nLite really bad, then it won't be easy to finish it. So as far, as validation go, I'm very skeptical. Win98 can boot easily, I have had this proven.

Solar flares happen all the time and the cosmic rays originate from events that happen milions and billions years ago. No way to know what happen, but my machine is entirly stable overclocked as well, as underclocked. So there is a good chance... The Aquamark run, well, there will be a validation problem, but they did not want latest CPU-Z and I found thread where I was complaining about crashes with this mobo there:
MSI 6340 HWiNFO freeze on sensors
...and since I cannot found CPU-Z 1.61.3, then I try the 1.61.4 and for safety with the SMBus=0 option at first, so the score should be produced. However it is running too slow, I starting to worry if it even make to the finish in time...!

Any luck finding CPU-Z 1.61.3...?

...

Also this thread: Gotta be a record for SuperPi 32M - Page 3
...catched my eye and I was wondering, if I can slow down regular hardware to produce 12+h SuperPi 32M runs:

Even on relatively modern hardware, like AthlonXP CPU's, most of these old board can disable L1 and L2 caches, some even the SSE instructions support. Even that disabling all but L1 does not affect the performance much, disabling L1 is "super-killer" of performance.

So that makes me wonder, how slow it will go on the SuperPi 32M test.

For the reference, all these machines can do SuperPi 32M w/o fail or crash or any other problems, all the computers have replaced their original bad caps to good ones and all these computers are 100% stable and I can run on th SuperPi 32M tests repeatedly w/o any trouble... BUT!

Here we come. I got starting scores 12h 28min per loop, or even 17h 59min per loop:





The later would produce (18hx24) a 432h long benchmark (18 days), and that is 1100MHz AXP using 256MB SDRAM (PCchips M810LR mobo).

However, and that is where things get really bad, it always crash. Sooner or later it crash on any machine with disabled L1 cache in bios I run it. A good example is P4 3.4GHz CPU that fail w/o caches too:



I would once again stress, that the machines are completely stable, there is no way that the error is hardware related. Not when it happen like 10x times.

...

So I stopped trying the cache and I added CPU stressing (for example latest CPU-Z v1.73) that run at high priority, while SuperPi do run at low priority, causing it to slow-down to point that the whole calculation of 32M will took about 18 to 24h, depending on settings of the priority and stuff.

Now imagine my surprise, when this also started and continued crashing. Numerous attepmts did not yield one SINGLE result. I'm "a bit" surprised and I cannot explain this behaviour. Normally systems w/o cache works just deadly slow (Aquamark3 is running on Duron 750 at 30x7.5 without caches since 13.8. and it is hardly in the half at the time of the posting - 27.8.), but works. No crashes or other problems (unless I click too much, one have to be carefull, even mouse have like 10sec lag when moved in Aquamark 3 :D ) happen, so my question is deadly simple - anyone have any idea, why this is happening?

Ano no, please don't say stability. These boards ARE reliable and stable. They can run SuperPi 32M each day releatedly, but not w/o caches or when slowed down, witch is weird at least.
__________________
"It is dangerous to be right in matters on which the established authorities are wrong." - Voltaire
"I believe that all the people who stand to profit by a war and who help provoke it should be shot on the first day it starts..." - Hemingway :) my config - my caps
Reply With Quote
  #64 (permalink)  
Old August 28, 2015, 03:42 AM
trodas's Avatar
MVP
 
Join Date: Feb 2008
Location: Czech republic
Posts: 394

My System Specs

Default

Oh, well... at least this is getting close to the end:



Who have fingers crossed...?
__________________
"It is dangerous to be right in matters on which the established authorities are wrong." - Voltaire
"I believe that all the people who stand to profit by a war and who help provoke it should be shot on the first day it starts..." - Hemingway :) my config - my caps
Reply With Quote
  #65 (permalink)  
Old August 28, 2015, 04:55 AM
Johan45's Avatar
Top Prospect
 
Join Date: Mar 2015
Location: South of Woodstock Ontario.
Posts: 171
Default

I don't know about disabled cache but I managed to make SPI32 run on a sempron at 50MHz. I was trying to use it to slow down AM3 but it didn't work.

__________________
Not sure if that's my wife's cooking or if I'm overclocking too much.
Reply With Quote
  #66 (permalink)  
Old August 28, 2015, 11:10 AM
trodas's Avatar
MVP
 
Join Date: Feb 2008
Location: Czech republic
Posts: 394

My System Specs

Default

These superlow multiplier on some AMD CPU's are supercool. I seen x0.25 multi... crazy. I slowed down Aquamark with:
Athlon 30MHz FSB x 9 + cache off. Aquamark is running since 13.8. for me and it does not made even 3k framers from the 5200 need. So, go figure... Hardly can anyone beat that, pls consider trying other challenges.

Also I would say that GPUPI is settled. 16 days, 8hours...
trodas`s GPUPI for CPU - 1B score: 16days 8h 45min 41sec 760ms with a Core 2 X6800 (2.93Ghz)

No more! :D
__________________
"It is dangerous to be right in matters on which the established authorities are wrong." - Voltaire
"I believe that all the people who stand to profit by a war and who help provoke it should be shot on the first day it starts..." - Hemingway :) my config - my caps
Reply With Quote
  #67 (permalink)  
Old August 29, 2015, 06:22 AM
Johan45's Avatar
Top Prospect
 
Join Date: Mar 2015
Location: South of Woodstock Ontario.
Posts: 171
Default

Nice work on GPUPi Trodas, that's a long haul. Last I looked at mine was 120 hrs for 3 iterations.
__________________
Not sure if that's my wife's cooking or if I'm overclocking too much.
Reply With Quote
  #68 (permalink)  
Old August 29, 2015, 06:35 AM
Dead Things's Avatar
Hall Of Fame
F@H
 
Join Date: Oct 2008
Location: Centre of the Universe
Posts: 1,969

My System Specs

Default

That's a well-laid gauntlet trodas! Nicely done.
__________________
Think you can overclock? Then show us what you got!
Join the Hardware Canucks Overclocking team today!

Follow my benching, folding, mining and miscellaneous shenanigans @dt_oc
Reply With Quote
  #69 (permalink)  
Old August 29, 2015, 09:19 AM
trodas's Avatar
MVP
 
Join Date: Feb 2008
Location: Czech republic
Posts: 394

My System Specs

Default

Dead Things - Thank you! However it looks like that I need help - there are people who claim lack of validation on the score, when _mat_ alone (autor of GPUPI) say that the file generated IS valid:


Help...! That is unfair :( I worked so hard and the video prove, that there ARE refresh issues, so taking the score down on such cheap example... that would be devastating for me... :( Pls help!

I can fight that BS alone... besides I'm busy by figuring out how to produce verifiable scores for the low CPU clocks on my Asus TXP4-X mobo (Intel could be 7.14, AMD could be 10.7MHz - that should have show them, who is the boss, right?)...




Johan45 - 120h for there loops? You sure you manage to do it till 9.30.2015? That is because if your average loop is 40h (120/3), then you get about 800h time, witch is 33.333 days. Not good, you did not have that much time... Futhermore my first there loops are done in 8h:
GPUPI_longer_runs_now_zpsgjw8opau.jpg Photo by ax2cz | Photobucket
...and then, since loop 6 or so it get _really_ slow w/o any change on MY side. So I quess that you get the lowest GPUPI score, but just not manage the deadline.

And since the cheap attack aimed at taking down my score might (w/o your help!) succeed, then we lose that run completely... :(
__________________
"It is dangerous to be right in matters on which the established authorities are wrong." - Voltaire
"I believe that all the people who stand to profit by a war and who help provoke it should be shot on the first day it starts..." - Hemingway :) my config - my caps
Reply With Quote
  #70 (permalink)  
Old August 29, 2015, 09:16 PM
Dead Things's Avatar
Hall Of Fame
F@H
 
Join Date: Oct 2008
Location: Centre of the Universe
Posts: 1,969

My System Specs

Default

I've got your back trodas.
__________________
Think you can overclock? Then show us what you got!
Join the Hardware Canucks Overclocking team today!

Follow my benching, folding, mining and miscellaneous shenanigans @dt_oc
Reply With Quote
Reply


Thread Tools
Display Modes

Similar Threads
Thread Thread Starter Forum Replies Last Post
Support Hardware Canucks Movember team!! stoanee Off Topic 56 December 1, 2012 08:31 PM
Robscix Leaves Hardware Canucks Review team Robscix Audio 14 March 25, 2009 03:12 PM
hardware canucks HWBot team? lowfat Overclocking, Tweaking and Benchmarking 105 February 13, 2008 02:22 AM
Hardware Canucks team Stats links sswilson HardwareCanucks F@H Team 0 May 20, 2007 12:06 AM