GPU FLOPS: Theory vs Reality

Message boards : Number crunching : GPU FLOPS: Theory vs Reality
Message board moderation

To post messages, you must log in.

Previous · 1 . . . 9 · 10 · 11 · 12 · 13 · 14 · 15 . . . 20 · Next

AuthorMessage
Profile Bill Special Project $75 donor
Volunteer tester
Avatar

Send message
Joined: 30 Nov 05
Posts: 282
Credit: 6,916,194
RAC: 60
United States
Message 1985823 - Posted: 18 Mar 2019, 17:09:19 UTC - in response to Message 1984636.  

I just noticed that the bars appear to be wider than the last graph. Any idea what might be the cause of it? Perhaps because more tasks have been crunched since the last plot?
Seti@home classic: 1,456 results, 1.613 years CPU time
ID: 1985823 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13161
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1985836 - Posted: 18 Mar 2019, 17:59:40 UTC - in response to Message 1985823.  

Believe the bars show variance. So larger sample size.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1985836 · Report as offensive
Profile Shaggie76
Avatar

Send message
Joined: 9 Oct 09
Posts: 282
Credit: 271,858,118
RAC: 196
Canada
Message 1988037 - Posted: 30 Mar 2019, 20:25:01 UTC

New scan includes early results for the GTX 1660 Ti -- only 14 hosts so there's more variance but it's looking very good.

ID: 1988037 · Report as offensive
Profile petri33
Volunteer tester

Send message
Joined: 6 Jun 02
Posts: 1668
Credit: 623,086,772
RAC: 156
Finland
Message 1988042 - Posted: 30 Mar 2019, 21:05:05 UTC

Nice graphs,
Thanks.

The top #2 is doing about 6000 cr/hour per each one RTX2080Ti.
I just wish those easy to calculate numbers would be plotted with red X on the chart too. My own (#3) mix of 2 GTX1080, one 1080Ti and a TITAN V is a hard one to plot so I'm not asking to do that.

I know that plotting bigger numbers to the graph would need a huge readjustment of the horizontal axis.

Thanks anyway.

I'd really like blue/green/gold bars to show NV special app performance per hour and per advertised watt for each GPU type that have any data in the DB. That way people could decide what to buy for Windows and what to buy when planning to build a (compute only?) Linux rig.

--
Petri
To overcome Heisenbergs:
"You can't always get what you want / but if you try sometimes you just might find / you get what you need." -- Rolling Stones
ID: 1988042 · Report as offensive
Profile Tom M
Volunteer tester

Send message
Joined: 28 Nov 02
Posts: 5124
Credit: 276,046,078
RAC: 462
Message 1988059 - Posted: 30 Mar 2019, 22:28:48 UTC - in response to Message 1988042.  


I'd really like blue/green/gold bars to show NV special app performance per hour and per advertised watt for each GPU type that have any data in the DB. That way people could decide what to buy for Windows and what to buy when planning to build a (compute only?) Linux rig.

--
Petri


I think Shaggy can't automate recognition of the CUDA special sauce.

I would be willing to create a text file of system ids of "all" or at least a lot of the special sauce users. Surely we have enough running them to get a good baseline. Then the trick would be to clone/create a new script that extracted only the systems on "the list".

Tom
A proud member of the OFA (Old Farts Association).
ID: 1988059 · Report as offensive
Grant (SSSF)
Volunteer tester

Send message
Joined: 19 Aug 99
Posts: 13720
Credit: 208,696,464
RAC: 304
Australia
Message 1988062 - Posted: 30 Mar 2019, 22:31:28 UTC - in response to Message 1988037.  

New scan includes early results for the GTX 1660 Ti -- only 14 hosts so there's more variance but it's looking very good.

Basically on par with the GTX 1080, and the RTX 2060 is on par with the GTX 1080Ti. Very good indeed.
Grant
Darwin NT
ID: 1988062 · Report as offensive
Profile Tom M
Volunteer tester

Send message
Joined: 28 Nov 02
Posts: 5124
Credit: 276,046,078
RAC: 462
Message 1988067 - Posted: 30 Mar 2019, 22:37:04 UTC - in response to Message 1988062.  

New scan includes early results for the GTX 1660 Ti -- only 14 hosts so there's more variance but it's looking very good.

Basically on par with the GTX 1080, and the RTX 2060 is on par with the GTX 1080Ti. Very good indeed.


I haven't looked. Is the price of a Gtx 1660 Ti at or lower than the price of a used gtx 1080 Ti? I am ignoring operational cost.

Tom
A proud member of the OFA (Old Farts Association).
ID: 1988067 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13161
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1988071 - Posted: 30 Mar 2019, 23:01:57 UTC - in response to Message 1988067.  

I haven't looked. Is the price of a Gtx 1660 Ti at or lower than the price of a used gtx 1080 Ti? I am ignoring operational cost.

Based on a quick perusal of used eBay prices of the 1080 Ti, the 1660 TI is half the price.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1988071 · Report as offensive
Ian&Steve C.
Avatar

Send message
Joined: 28 Sep 99
Posts: 4267
Credit: 1,282,604,591
RAC: 6,640
United States
Message 1988073 - Posted: 30 Mar 2019, 23:16:51 UTC - in response to Message 1988062.  

New scan includes early results for the GTX 1660 Ti -- only 14 hosts so there's more variance but it's looking very good.

Basically on par with the GTX 1080, and the RTX 2060 is on par with the GTX 1080Ti. Very good indeed.


there's too much variance with the stock apps due to all the different settings people run.

i've compared several models manually by watching the performance of many hosts with various cards with the special app, which has MUCH less variance since there's really only one switch you can play with (the nobs argument).

hierarchy is roughly this:
Titan V
2080ti
2080
2070 ~ 1080ti
2060 ~ 1080
1660ti ~ 1070ti
1660 likely to be around a 1070, but i haven't looked.

I'm in the process of swapping my 1080ti's to 2070s to save on some power. they are almost identical in speed.
Seti@Home classic workunits: 29,492 CPU time: 134,419 hours

ID: 1988073 · Report as offensive
Profile bloodrain
Volunteer tester
Avatar

Send message
Joined: 8 Dec 08
Posts: 231
Credit: 28,112,547
RAC: 1
Antarctica
Message 1988091 - Posted: 31 Mar 2019, 2:49:21 UTC - in response to Message 1891126.  

good to here. i got got 4 more and will add them to wu next month.
also i should note for some reason i say last 3 driver update from amd been doing some strange things to the 580s
ID: 1988091 · Report as offensive
Profile Shaggie76
Avatar

Send message
Joined: 9 Oct 09
Posts: 282
Credit: 271,858,118
RAC: 196
Canada
Message 1988200 - Posted: 31 Mar 2019, 20:37:08 UTC - in response to Message 1988042.  

I just wish those easy to calculate numbers would be plotted with red X on the chart too.

As a programmer I'm sure that you're familiar with Amdahl's law; making the app 3x faster for a dozen users is insignificant compared with making the app faster for the 90,000 other active users volunteering for the project. Please stop asking me to show how much faster your code is on my graphs until it's an official app that everyone can use.
ID: 1988200 · 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: 9786
Credit: 572,710,851
RAC: 3,799
Panama
Message 1988216 - Posted: 31 Mar 2019, 22:31:56 UTC - in response to Message 1988200.  
Last modified: 31 Mar 2019, 22:35:40 UTC

Please stop asking me to show how much faster your code is on my graphs until it's an official app that everyone can use.

I was one of those who ask, it's your app & your choice, we all need to respect that of course, just was unaware of that. Will not ask again in the future. Please forgive me.
Just not understand the second part, there are some misunderstanding, the Linux apps are available in the same site as the Lunatics are, so anyone could use.
ID: 1988216 · Report as offensive
Profile Tom M
Volunteer tester

Send message
Joined: 28 Nov 02
Posts: 5124
Credit: 276,046,078
RAC: 462
Message 1989215 - Posted: 7 Apr 2019, 22:36:26 UTC - in response to Message 1984636.  

New this scan: NVIDIA GeForce RTX 2060's with excellent credit/watt and excellent performance.

There were some GTX 1660 Ti's in the scan but not quiet enough to qualify. I'll re-scan in a few weeks and maybe there'll be enough by then.



I keep running across (gamer) reviews that claim a Gtx 1070 is so close to being a Gtx 1070Ti that you should buy the Gtx 1070. I know, on the used market at least, that there is $70- difference in price. Anyone have an opinion?

Tom
A proud member of the OFA (Old Farts Association).
ID: 1989215 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13161
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1989217 - Posted: 7 Apr 2019, 22:39:31 UTC - in response to Message 1989215.  

I keep running across (gamer) reviews that claim a Gtx 1070 is so close to being a Gtx 1070Ti that you should buy the Gtx 1070. I know, on the used market at least, that there is $70- difference in price. Anyone have an opinion?


I am always disappointed that my 1070Ti's performance is closer to the 1070 than it is to my 1080's.

Even though it only has one less SM compared to the 1080 and 4 more SM's than my 1070's
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1989217 · Report as offensive
Profile Tom M
Volunteer tester

Send message
Joined: 28 Nov 02
Posts: 5124
Credit: 276,046,078
RAC: 462
Message 1989456 - Posted: 10 Apr 2019, 12:39:20 UTC - in response to Message 1989217.  

I keep running across (gamer) reviews that claim a Gtx 1070 is so close to being a Gtx 1070Ti that you should buy the Gtx 1070. I know, on the used market at least, that there is $70- difference in price. Anyone have an opinion?


I am always disappointed that my 1070Ti's performance is closer to the 1070 than it is to my 1080's.

Even though it only has one less SM compared to the 1080 and 4 more SM's than my 1070's


The price of a gtx 1070 (used) vs. gtx 1070Ti on eBay is $80-$100 less. Are there any other markets where I can reliably get one (or more) cheaper?

Tom
A proud member of the OFA (Old Farts Association).
ID: 1989456 · Report as offensive
Profile Tom M
Volunteer tester

Send message
Joined: 28 Nov 02
Posts: 5124
Credit: 276,046,078
RAC: 462
Message 1989514 - Posted: 10 Apr 2019, 15:59:44 UTC - in response to Message 1989456.  


The price of a gtx 1070 (used) vs. gtx 1070Ti on eBay is $80-$100 less. Are there any other markets where I can reliably get one (or more) cheaper?

Tom


Just found some used gtx 1070's on eBay for $88, with a brand name (if they don't offer a brand beyond Nvidia, it is usually a scam offer), offered from the US. So I ordered 2.

Tom
A proud member of the OFA (Old Farts Association).
ID: 1989514 · Report as offensive
Profile Tom M
Volunteer tester

Send message
Joined: 28 Nov 02
Posts: 5124
Credit: 276,046,078
RAC: 462
Message 1989751 - Posted: 12 Apr 2019, 6:24:49 UTC - in response to Message 1989514.  


The price of a gtx 1070 (used) vs. gtx 1070Ti on eBay is $80-$100 less. Are there any other markets where I can reliably get one (or more) cheaper?

Tom


Just found some used gtx 1070's on eBay for $88, with a brand name (if they don't offer a brand beyond Nvidia, it is usually a scam offer), offered from the US. So I ordered 2.

Tom


So far it is looking like yet another dead end :(

Tom
A proud member of the OFA (Old Farts Association).
ID: 1989751 · Report as offensive
Profile Shaggie76
Avatar

Send message
Joined: 9 Oct 09
Posts: 282
Credit: 271,858,118
RAC: 196
Canada
Message 1997116 - Posted: 7 Jun 2019, 1:58:34 UTC

New this scan, Radeon RX 590, GeForce GTX 1660 Ti, and GeForce RTX 2060: the 1660 Ti looks to be a new leader in performance/watt and the 2060 looks like an excellent economy option!

ID: 1997116 · Report as offensive
.:krypt08700:.

Send message
Joined: 20 Jul 02
Posts: 5
Credit: 6,788,712
RAC: 0
United States
Message 1997117 - Posted: 7 Jun 2019, 2:22:42 UTC

Very cool information.
ID: 1997117 · Report as offensive
Profile Keith Myers Special Project $250 donor
Volunteer tester
Avatar

Send message
Joined: 29 Apr 01
Posts: 13161
Credit: 1,160,866,277
RAC: 1,873
United States
Message 1997120 - Posted: 7 Jun 2019, 2:54:39 UTC

Thanks for the updated charts and stats, Shaggie76. I have my sights set on the 2070 as the best compromise to replace the 1070's.
Seti@Home classic workunits:20,676 CPU time:74,226 hours

A proud member of the OFA (Old Farts Association)
ID: 1997120 · Report as offensive
Previous · 1 . . . 9 · 10 · 11 · 12 · 13 · 14 · 15 . . . 20 · Next

Message boards : Number crunching : GPU FLOPS: Theory vs Reality


 
©2024 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.