PrimeGrid Races 2018

[zzuupp]

Thanks for the stats, Ken.

& WOW! we are gaining on Seti

 

[StefanR5R]

This challenge is full of surprises.
First, ATP miss to put much of their pool on the job.
Then, [H]ard|OCP are cautiously taking a renewed interest in this PrimeGrid thing.
Still 28 hours to go, so what comes next?

 

[Ken g6]

Day 4 stats, late again:

Rank___Credits____Username
1______4843112____xii5ku
6______2338909____emoga
10_____1723458____crashtech
26_____1014470___10esseeTony
33_____726178_____Tejas
37_____687944_____biodoc
44_____601342_____markfw
47_____553396_____Howdy2u2
65_____386023_____Ken_g6
101____244158_____Orange Kid
121____198548_____zzuupp
178____111293_____Kiska
235____53642______indydude, Son of crashtech
247____45718______SlangNRox

Rank__Credits____Team
1_____16671175___Czech National Team
2_____13970764___SETI.Germany
3_____13528197___TeAm AnandTech
4_____11873381___Aggie The Pew
5_____8785172____Sicituradastra.
6_____3293692____BOINC@AUSTRALIA

We're close to SG, but they're holding a pretty constant lead on us. Still, third is the best we've ever done.

 

[Markfw]

Day 4 stats, late again:

Rank___Credits____Username
1______4843112____xii5ku
6______2338909____emoga
10_____1723458____crashtech
26_____1014470___10esseeTony
33_____726178_____Tejas
37_____687944_____biodoc
44_____601342_____markfw
47_____553396_____Howdy2u2
65_____386023_____Ken_g6
101____244158_____Orange Kid
121____198548_____zzuupp
178____111293_____Kiska
235____53642______indydude, Son of crashtech
247____45718______SlangNRox

Rank__Credits____Team
1_____16671175___Czech National Team
2_____13970764___SETI.Germany
3_____13528197___TeAm AnandTech
4_____11873381___Aggie The Pew
5_____8785172____Sicituradastra.
6_____3293692____BOINC@AUSTRALIA

We're close to SG, but they're holding a pretty constant lead on us. Still, third is the best we've ever done.

Do you thing another 56 threads would make a difference ? When I added 100 I see the jump.

 

[crashtech]

Do you thing another 56 threads would make a difference ? When I added 100 I see the jump.

I think it might, it would be worth it just to see SG panic!

 

[Tejas_2]

Thanks for the stats.

 

[Markfw]

This is my app_config.xml
Don't know if the "<app></app>" part is needed, but it works for me.

<app_config>
    <app>
        <name>llrSR5</name>
       <max_concurrent>2</max_concurrent>
    </app>
    <app_version>
        <app_name>llrSR5</app_name>
        <cmdline>-t 8</cmdline>
        <avg_ncpus>8</avg_ncpus>
    </app_version>
</app_config>

Where is app_config in linux ?

 

[Markfw]

well, it says the tasks will take 2 days (x 32 threads) so until I know where to put it, I can't help our race

 

[Ken g6]

Where is app_config in linux ?

/var/lib/boinc-client/projects/www.primegrid.com is where you put an app_config.xml

 

[Markfw]

/var/lib/boinc-client/projects/www.primegrid.com is where you put an app_config.xml

well, I am not the owner ?? and I can't do anything with this directory

Edit: if I right click and open as root (using the same password as my user) I can get in...

linux is so odd

Edit 2: so I set it to 8, and under Windows thats 8 threads. In linux with 32 threads and 8 cores, it thinks I can only do 2 8 threads clients ???

 

[Markfw]

OK, I have tried everything, but I can't get it to use all 32 threads. What do I set it to ??? Current is : (I have tried 4,8,16 for each line to no avail. Under windows 4,4 works great

<app_config>
<app_version>
<app_name>llrSR5</app_name>
<cmdline>-t 4</cmdline>
<avg_ncpus>4</avg_ncpus>
</app_version>
</app_config>

 

[StefanR5R]

The app_cofing.xml looks good to me. After you saved app_config.xml, remember to shutdown and restart the client. On Linux Mint, the command is

sudo /etc/init.d/boinc-client restart​

Or reboot the entire host.

Or set "Leave non-GPU config tasks in memory while suspended" to off, then suspend computation, use "Options/ Read config files", and finally resume computation.

 

[crashtech]

You might try setting it to seven, this will get as close to full utilization as possible while leaving some resources for the system. I've noticed that trying to allocate all cores to SR5 is in some cases counterproductive because it's such a demanding app. Might be worth a try, anyway. Five or six would work as well and get closer to full utilization, but then each discrete task will not be confined to its own core, which may or may not cause a performance hit.

 

[Kiska]

There is another way of seeing runtime, you can look inside the boinc dir and find a file named: "job_log_www.primegrid.com.txt" that should tell you pretty much everything

 

[Kiska]

Using mine as an example:

1529318087 ue 42829.599467 ct 45654.906200 fe 137777876931160 nm llrSR5_299591380_1 et 6915.924587 es 0

I believe the format is

(Epoch time) ue (estimated runtime uncorrected(boinc interpretation of runtime)) ct (final CPU time) fe (FLOPs estimate) nm (WU name) et (final elapsed time) es (exit status)

Using that we can interpret:
We can see it was submitted on Monday, June 18, 2018 11:18:14 AM GMT
BOINC's scheduler estimated 42829.59...... seconds of runtime
The actual CPU runtime was 45654.90 seconds
This task consumed 137.7779 TFLOPs of compute
The task name was llrSR5_299591380_1 therefore the WU name is llrSR5_299591380
It took 6915 seconds to complete this task
And the exit status was 0, which means its ok

 

[StefanR5R]

If I get my new 2700X done tonight, I will use it as a OC test....

One Ryzen has 66 % success rate,
the other one 0 % at this time. PrimeGrid is very sensitive once you go past stock clocks.

 

[ericlp]

Ignore please.... whoops, wrong thread... too early....

 

[Markfw]

One Ryzen has 66 % success rate,
the other one 0 % at this time. PrimeGrid is very sensitive once you go past stock clocks.

The stock one has 0%. The overclocked one is 66%

 

[StefanR5R]

The stock one has 0%.

Hmm. Is the memory faulty? But then I would expect only a few tasks getting invalidated, not all. Or is the processor voltage below stock?

 

[biodoc]

Using mine as an example:

1529318087 ue 42829.599467 ct 45654.906200 fe 137777876931160 nm llrSR5_299591380_1 et 6915.924587 es 0

I believe the format is

(Epoch time) ue (estimated runtime uncorrected(boinc interpretation of runtime)) ct (final CPU time) fe (FLOPs estimate) nm (WU name) et (final elapsed time) es (exit status)

Using that we can interpret:
We can see it was submitted on Monday, June 18, 2018 11:18:14 AM GMT
BOINC's scheduler estimated 42829.59...... seconds of runtime
The actual CPU runtime was 45654.90 seconds
This task consumed 137.7779 TFLOPs of compute
The task name was llrSR5_299591380_1 therefore the WU name is llrSR5_299591380
It took 6915 seconds to complete this task
And the exit status was 0, which means its ok

Thanks Kiska, this is useful. The run times in this file are consistent with from what I have observed when downloading a bunch of tasks, turning off networking and then recording run times of completed task ready to upload. The run times are still not consistent with run times reported by the primegrid server.

I did figure out what the server run times are though and you are not going to believe it.

It's simply the difference between server time sent and time reported. It's a number that's useless to us.

 

[Markfw]

Hmm. Is the memory faulty? But then I would expect only a few tasks getting invalidated, not all. Or is the processor voltage below stock?

The memory is rated at 4266, and only running at 3600. But I set it down to 3466 after you posted that, as the motherboard may not handle anything faster than 3466.

 

[TennesseeTony]

I just noticed EVGA's bcavnaugh is waaaaaaayyyyyyy down at the bottom of this race. @yodap @bill1024 Is he alright? Not at all like him to skimp out on a PG race.....

 

[Markfw]

No stats yet for the end of the race ? it is over, isn't it ?

 

[Ken g6]

No stats yet for the end of the race ? it is over, isn't it ?

Yep, it's all over but the stats reporting and the double-checking. I'll get to it eventually.

Edit: Here's the basic stats:

Rank___Credits____Usern
me
1______6035175____xii5ku
7______2912829____emoga
10_____2196663____crashtech
20_____1501675___10esseeTony
35_____925383_____Tejas
36_____892216_____biodoc
40_____853093_____markfw
45_____763200_____Howdy2u2
72_____479317_____Ken_g6
104____301695_____Orange Kid
126____251564_____zzuupp
176____158409_____Kiska
249____59083______SlangNRox
252____56909______indydude, Son of crashtech

Rank__Credits____Team
1_____23839348___Czech National Team
2_____17699037___SETI.Germany
3_____17387217___TeAm AnandTech
4_____14603213___Aggie The Pew
5_____11093095___Sicituradastra.
6_____4055877____BOINC@Poland

SG was just a little too fast for us. But third is very nice.

 

[Orange Kid]

Thanks for the stats.
Third has got to be a first for us.!!!!