The Scottish Boinc Team

Date: 24-Jan-2020, 00:00:00 UTC to 25-Jan-2020 00:00:00 UTC
Project: 321-sieve (CPU only)

As with all PG Challenges, only workunits downloaded after the start of the challenge and completed and returned by the end of the challenge are valid (double-check validation not required).

This is the first of nine challenges this year, the only 1-day challenge in the series ( http://www.primegrid.com/challenge/challenge.php ), and the only one till mid-March.

Workunits are quick ( under 10 min on EYPC 7281 @ 2.1ghz, under 3 min on i9-9900K @ 3.6ghz edit: wrong sieve / wrong times/ right clown. corrected: 1hr 10min - EPYC, 50 min - i9 ) and are best run with hyper-threading on ( as opposed to most PG cpu apps which use the LLR algorithm and prefer HT off ). edit: 609.89 pts / wu

Please feel free to join me in this inaugural challenge,
g

Perfect. I need less than 5k hours to get WUProp lime 25k hours on that app.

Dipping a toe in the water now. Any reason not to run all available threads? That is, is there a minimum RAM or L3 Cache bottleneck that you know of on larger machines?

ETA : low RAM requirement (137M on Linux)

Gibson Praise wrote: ↑Sat Jan 18, 2020 2:43 am 50 min - i9

Yup - same on 3990X https://tsbt.co.uk/forum/viewtopic.php? ... 758#p79758

reminder - starts at midnight UTC

Ready!

For maximum credit during the challenge, clear your cache before the start and don't download any new WU until after the start. It's only a 24 hour challenge so it will make a difference.

Time to go!

I am fully loaded. About 20mins until first WUs turned in

Somebody get a picture ... we're in 3rd.

davidBAM wrote: ↑Fri Jan 24, 2020 1:57 am

.. and still in third
and on the individual side davidBAM continues to hold first with Bowmore close behind

Hmm ... no denying first place individual be nice. Maybe some of the Pros will just hire clout from Amazon Web Services though?

Third placed team will ruffle feathers

There is an entire thread devoted to Google CoLab http://www.primegrid.com/forum_thread.p ... 839#135424 .. mostly devoted to GPU use. Haven't seen anyone try to abuse the CPUs yet.

On a one day challenge I think we might catch them by surprise.

edit: Woohoo! I am back into the top 50 - which for me is fabulous!

I'm starting to get run over! We're still holding 3rd though.

Bryan wrote: ↑Fri Jan 24, 2020 6:56 am I'm starting to get run over!

Me too

We are about to drop to 5th. Any assistance greatly appreciated

Hmmm I'm running 321-sieve but I don't seem to be showing on the Challenge page for the team (or as an individual for that matter).

Checked on the Primegrid Account Page and I am deffo showing as TSBT.

According to BAM (not David) I have done 195,773 so far today.

@jockmacmad2 You are #66 https://www.primegrid.com/challenge/202 ... users.html

We dropped to 5th - now back at 4th. I've even got 8WU running on my laptop but I doubt if that is the reason.

Laptop crunching not recommended, I know

If you can crunch on the laptop, well I just started an old i7-3770. Might get ten or twenty workunits.

We are just a tad competitive on here

The laptop actually does quite well, 8 threads and 80mins each so 1 WU every 10mins

davidBAM wrote: ↑Fri Jan 24, 2020 6:39 pm We are just a tad competitive on here

Speak for yourself, there isn't a competitive bone in my body

NVM

NVM

nvm

NVM

NVM

My laptop went crazy! Sorry.

davidBAM
Bowmore
scole of TSBT
Gibson Praise
JockMacMad TSBT
Alez

Well done all, 5th in the Team Place which is awesome against the PrimeGrid Professionals. We had 3 in the Top 20, 5 in the Top 100.

I got first individual place with more than a little help from Bryan who sacrificed a Top 5 finish for himself to hep me in last 7-8 hours - many, many thanks. I'd been having a ding-dong battle staying just ahead of RFGuy from [H] team so Bryan stepped in to make sure of things

Heckuva showing. 4th was tantalizingly close!

We could have made 3rd with a concerted effort, I reckon.

At risk of blowing my own trumpet, I have to point out how much btsk helped me - I had 480 threads running flat-out in 24 instances on 18 machines within 9-10mins of the start. btsk ticklers also ensured all had a full day's work within 30mins (and it did require a bit of tickling even though PG is an extremely well-run project. Okay, so Bryan helped a bit at the end but it really was a flying start