128 Thread chip??

Message boards : Number crunching : 128 Thread chip??
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile Zalster Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 27 May 99
Posts: 4803
Credit: 317,387,132
RAC: 245,737
United States
Message 1920041 - Posted: 20 Feb 2018, 1:59:59 UTC

ID: 1920041 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 6239
Credit: 446,903,636
RAC: 1,078,097
United States
Message 1920060 - Posted: 20 Feb 2018, 3:36:35 UTC - in response to Message 1920041.  

I believe the "Rome"processor was on the cpu timeline PowerPoint slides at the inaugural Ryzen launch. So nothing unexpected. I have gathered that the TSMC foundry 7nm fab line is further along that many thought. The thinking though was that Qualcomm and Apple would have first shot at production relegating AMD to fill in the minimal production gap slots. Therefore, the "Rome" processors wouldn't be available in any quantities other than engineering samples to whet the datacenter purchasing managers appetites in late 2018. Maybe some slight product availability in 2019. Question is who will get priority at AMD, Zen2 or Navi GPU's for the 7nm parts. Does AMD think there is greater chance to grab market share in gpus or cpus?

Granted a 128 thread cpu boggles the mind.
Seti@Home classic workunits:20,676 CPU time:74,226 hours
ID: 1920060 · Report as offensive
Profile Brent Norman Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester

Send message
Joined: 1 Dec 99
Posts: 2547
Credit: 314,245,229
RAC: 740,294
Canada
Message 1920065 - Posted: 20 Feb 2018, 4:14:45 UTC

My thought is ... SETI has a 100 task limit per device and couldn't even feed it, LOL.
ID: 1920065 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 6239
Credit: 446,903,636
RAC: 1,078,097
United States
Message 1920067 - Posted: 20 Feb 2018, 4:22:34 UTC - in response to Message 1920065.  

My thought is ... SETI has a 100 task limit per device and couldn't even feed it, LOL.

Now that is painfully funny, LOL. Wish we could get the BOINC developers to come into the modern age.
Seti@Home classic workunits:20,676 CPU time:74,226 hours
ID: 1920067 · Report as offensive
juan BFP Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 16 Mar 07
Posts: 7226
Credit: 424,659,622
RAC: 234,906
Panama
Message 1920070 - Posted: 20 Feb 2018, 4:40:00 UTC
Last modified: 20 Feb 2018, 4:42:09 UTC

Hide your credit cards. Just imagine the price of each one of this toys.

A simple math... 128 (threads) x 1 hr (crunch time per WU) x 24 hrs = > 3000 WU /Day per CPU... Something like > 200K/day even with Creditscrew. If is possible to keep it feeded of course.
ID: 1920070 · Report as offensive
Profile betreger Project Donor
Avatar

Send message
Joined: 29 Jun 99
Posts: 8531
Credit: 20,397,173
RAC: 12,824
United States
Message 1920071 - Posted: 20 Feb 2018, 4:51:51 UTC - in response to Message 1920067.  

My thought is ... SETI has a 100 task limit per device and couldn't even feed it, LOL.

Now that is painfully funny, LOL. Wish we could get the BOINC developers to come into the modern age.

You know the server is the cause of that. The classic fix is throw a bigger gorilla at it.
ID: 1920071 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 6239
Credit: 446,903,636
RAC: 1,078,097
United States
Message 1920081 - Posted: 20 Feb 2018, 6:43:18 UTC - in response to Message 1920070.  

Hide your credit cards. Just imagine the price of each one of this toys.

A simple math... 128 (threads) x 1 hr (crunch time per WU) x 24 hrs = > 3000 WU /Day per CPU... Something like > 200K/day even with Creditscrew. If is possible to keep it feeded of course.

You need to rework your math. In my case . . . 128 threads X 30minutes (crunch time per WU and SSE41 app) X 24 hours => 6144 WU /Day per CPU >> 307K/day at 50 credits per WU with CreditScrew.
Seti@Home classic workunits:20,676 CPU time:74,226 hours
ID: 1920081 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 6239
Credit: 446,903,636
RAC: 1,078,097
United States
Message 1920082 - Posted: 20 Feb 2018, 6:44:42 UTC - in response to Message 1920071.  

My thought is ... SETI has a 100 task limit per device and couldn't even feed it, LOL.

Now that is painfully funny, LOL. Wish we could get the BOINC developers to come into the modern age.

You know the server is the cause of that. The classic fix is throw a bigger gorilla at it.

But the developers write the scheduler code.
Seti@Home classic workunits:20,676 CPU time:74,226 hours
ID: 1920082 · Report as offensive
juan BFP Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 16 Mar 07
Posts: 7226
Credit: 424,659,622
RAC: 234,906
Panama
Message 1920084 - Posted: 20 Feb 2018, 7:34:00 UTC - in response to Message 1920081.  

Hide your credit cards. Just imagine the price of each one of this toys.

A simple math... 128 (threads) x 1 hr (crunch time per WU) x 24 hrs = > 3000 WU /Day per CPU... Something like > 200K/day even with Creditscrew. If is possible to keep it feeded of course.

You need to rework your math. In my case . . . 128 threads X 30minutes (crunch time per WU and SSE41 app) X 24 hours => 6144 WU /Day per CPU >> 307K/day at 50 credits per WU with CreditScrew.

I was very very conservative...... LOL
ID: 1920084 · Report as offensive
Profile Chris Feyen
Volunteer tester

Send message
Joined: 1 Jan 01
Posts: 76
Credit: 17,196,356
RAC: 42,553
United States
Message 1920117 - Posted: 21 Feb 2018, 1:00:39 UTC - in response to Message 1920065.  

My thought is ... SETI has a 100 task limit per device and couldn't even feed it, LOL.


That's kinda funny, in a sad way.
~Chris

ID: 1920117 · Report as offensive
Al Crowdfunding Project Donor*Special Project $75 donorSpecial Project $250 donor
Avatar

Send message
Joined: 3 Apr 99
Posts: 1676
Credit: 397,308,376
RAC: 312,245
United States
Message 1920125 - Posted: 21 Feb 2018, 1:30:36 UTC

Welll! Look at that won't you.. I bet its clocked relatively low, that seems to be the rule, the more cores, the slower the GHz. It'll be interesting when I get my 44/88 core system up and running (RedHat. Yeah, I know, I know, getting there...) but that wouldn't hold a candle to this 128 core monster. Any guesses as to the cost per chip? Imagine 2 of those on a server board.. Redonkulous! SETI wouldn't know what to do with it! lol

ID: 1920125 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 6239
Credit: 446,903,636
RAC: 1,078,097
United States
Message 1920127 - Posted: 21 Feb 2018, 1:56:22 UTC - in response to Message 1920125.  

Haven't a clue. What package? How many dies per package? If it follows Epyc layout, it would have 4 full active dies. Will they keep the 4094 contacts of the SP3r2 socket? Or have to add more and new socket density?
Seti@Home classic workunits:20,676 CPU time:74,226 hours
ID: 1920127 · Report as offensive
Profile betreger Project Donor
Avatar

Send message
Joined: 29 Jun 99
Posts: 8531
Credit: 20,397,173
RAC: 12,824
United States
Message 1920146 - Posted: 21 Feb 2018, 3:30:01 UTC - in response to Message 1920140.  

Oh
ID: 1920146 · Report as offensive
account
Volunteer tester

Send message
Joined: 31 Dec 99
Posts: 1716
Credit: 6,108,249
RAC: 3
Saint Pierre and Miquelon
Message 1920152 - Posted: 21 Feb 2018, 3:53:23 UTC

I'm pretty sure it's a drop in replacement for current epyc server chips.
Would probably need a bios upgrade.
ID: 1920152 · Report as offensive

Message boards : Number crunching : 128 Thread chip??


 
©2018 University of California
 
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.