Jump to content
HWBOT Community Forums

The official HWBOT Team Cup 2015 - SC4: MISC Challenge thread.


rtsurfer

Recommended Posts

Dunno why. Maybe it is just because people are trying hard to get Skylake into these benches?

Dunno. I'm satisfacted with my old hardware for *some* benches :)

 

But recently it seems to run too fast to me:

GPUPI_565h_batch_11.jpg

 

39h per batch is too fast :D :D :D

 

...

 

But it looks like to me, that I cannot beat Spiedie 946h score because of time reasons. He got 946h, and I'm at 565h now. 381h short of his time. 381h is 15.8 days and we are at 23. September, witch give only 7 days to deadline.

 

Also 9 remaining batches for 39h translate to "only" 351h, so I will be 30h short of the 946h record long time.

 

The question is only... should I keep the bench running to the end? :confused:

Edited by trodas
Link to comment
Share on other sites

  • Replies 76
  • Created
  • Last Reply

Top Posters In This Topic

You finally did the math i see :P

 

All my other slow runs have followed a 3x fast + 2x slow pattern, to the point that runtime = 27.5xfirst batch pretty accurately (or 27.5x any 5th batch, so batch 1, 6 ,11, 16, so 36h in the 946h run = 27.5x990h, close enough). Yours should then be 39x27.5 = 1072h. My plan of a 40 day run worked out pretty accurately i think :)

Link to comment
Share on other sites

Well, yea :) I should done it sooner, but I'm more after the refresh bug with this run. Previous attempt stabilized the refresh in the batch 12: http://img.hwbot.org/u8592/image_id_1475307.jpg

 

And my run was like 4 fast runs, very closely to each other and then it get haywire - calculated time just by the hours, so, not very precise (23:57 is still counted as 23, so...):

 

1 - 2h

2 - 3h

3 - 3h

4 - 3h

5 - 14h

6 - 23h

7 - 19h

8 - 26h

9 - 26h

10 - 28h

11 - 21h

12 - 24h

13 - 23h

14 - 23h

15 - 22h

16 - 21h

17 - 22h

18 - 21h

19 - 22h

20 - 22h

 

But it still are quire stabilized in the last half. I used two instances of CPU-Z for CPU stress, opened windows. Not 3 instances, but minimized. Win7 not activated and during the test it start complain about not genuine... So I keep it this way, might be relevant to the refresh bug...

 

Still the question is same - keep it running or stop?

Link to comment
Share on other sites

Sure I can kill the CPU-Z processes... but that cannot give me to win 1st place, not to mention the results with most of the batches invisible will be contraversial. Sure _mat_ will check the file and it will pass as good, because nothing wrong was done there, but... people will still complain. I can made it to last about 730h top, then it will be time to speed-up the calc and hurry to submit it.

 

What that could give...?

Link to comment
Share on other sites

Sure I can kill the CPU-Z processes... but that cannot give me to win 1st place, not to mention the results with most of the batches invisible will be contraversial. Sure _mat_ will check the file and it will pass as good, because nothing wrong was done there, but... people will still complain. I can made it to last about 730h top, then it will be time to speed-up the calc and hurry to submit it.

 

What that could give...?

 

Well, a second or third place is not a win but still a good place to finish in.

And yea Im still a tad grumpy that my GF killed my 700h run (at batch 14) when I was away ;)

Link to comment
Share on other sites

I will read this text (translate) to my GF, because she managed to avoid killing my Aquamark almost month run to score 2, just to kick the power line by carelessly walking by, when I was saving the score logs and pictures to my flashdrive under normal settings (at lowered FSB 34MHz the USB does not work)... Unbelievable.

 

So, since I having hard time to orientate on the site, then what is the score I need to get to second or third place? Provided that no-one run like me, to submit at the last possible moment...

Link to comment
Share on other sites

That could be megaslow, lol. (If he uses v2.1, then that could be done by ughly nasty trick too (cheat), I stumbled upon it and reported it...) But the problem is, that he cannot finish that in time :)

 

Mine are at 717h now, loop 14:

 

GPUPI_v2_2_717h_batch_14.jpg

 

:) It is probably time to stop the CPU load now... or...?

 

 

PS. so I figured the oc-esports.io page and the scores are as follows:

 

http://oc-esports.io/#!/round/team_cup_2015_sc4/2352/slowest_gpupi_for_cpu_1b

 

1st - 950h 11m

2nd - 946h 3m

3rd - 609h 25m

4th - 368h 45m

 

...while that 4th submission is mine and I can be on 3rd place, as there is simply not enought time for me to get into 1st :(

 

Now there are only two questions:

 

- how long I should keep it running so there is safe time that it will end before the end of 30 September day

 

- what should I do to get around the refresh problem, so the score image will not looks crazy as before... :rolleyes: I will take a vid from it (when someone feel the need to complain), but I cannot influence the screenshot into the resulting file :rolleyes:

 

 

And THIS is not looking good at all:

GPUPI_2_1_2_368h_45m_run.jpg

 

Maybe I could take my own screenshot and _mat_ or someone could replace the GPUPI internal screenshot it with in case it get "weird" again? :confused:

Edited by trodas
Link to comment
Share on other sites

BassPlayer was exaggerating slightly, only 420 hour loops. I don't believe its a cheat, just testing revealed the benchmarks Achilles heel but we will see if it will validate or not in the next day or so when I turn off the slowing process. If we are doing this stage for next year's Team Cup can we get the announcement now? With 420 hour loops if I start right now, I can just squeak in the entire benchmark in a year.

 

And I need a UPS.

 

Trodas that screenshot doesn't look valid at all, will that pass moderation?

Link to comment
Share on other sites

just testing revealed the benchmarks Achilles heel

 

Care to explain this? I use CPU-Z to slow things down, but it produce the refresh bug...

 

If we are doing this stage for next year's Team Cup can we get the announcement now? With 420 hour loops if I start right now, I can just squeak in the entire benchmark in a year

 

LOL ... that is a good point :D

 

Trodas that screenshot doesn't look valid at all, will that pass moderation?

 

That is refresh bug, that is there from the very beginning for me. Maybe someone with AGP graphic could confirm that some day, but the screenshot is - believe or not - valid. The computation file is valid, witch is what matters. The screenshot is enbeded with the benchmark, witch is why it is and it will stays that way. I have even a video that show at least two serious refreshing problems (whole remaining times disappear) from saving of the score (

), so it is valid. I'm not responsible for program bugses and quirks. Hardly I can do more that I already did... The machine is able to pass many benches, like PCmark 07: http://hwbot.org/submission/2945615_

Or GPUPI 1B: http://hwbot.org/submission/2944654_

 

So, I probably try making own screenshot before things disappear :)))

Link to comment
Share on other sites

No, just stoping of the CPU load... intesting for some, I believe. Some of the refreshbugs caused by not responding CPU-Z are visible...

 

And of course, in the end score, there is the refreshbug again!

http://img.hwbot.org/u8592/image_id_1496581.jpg

 

And the score cannot be submited - just now. Just great!

 

The GPUPi stage ended like 20 hours ago or something so if your trying to submit not you missed the deadline :(

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


×
×
  • Create New...