All things Nvidia GPU

Skip Da Shu
Skip Da Shu
Joined: 18 Jan 05
Posts: 157
Credit: 1135046820
RAC: 1538450

I have a Gigabyte 1660ti

I have a Gigabyte 1660ti listed on local FB marketplace:

https://www.facebook.com/marketplace/item/1565325204288928/?ref=category_feed&referral_code=null&referral_story_type=post&tracking=browse_serp%3A1d7c8c0f-227e-42da-bef6-7b8bc5cbe3fb

 

Also have a Gigabyte 1080ti I bought in December that I MIGHT replace later this month. N108TGAMINGOC-11GD

KLiK
KLiK
Joined: 1 Apr 14
Posts: 92
Credit: 517036365
RAC: 1136596

Phil wrote: Two computers

Phil wrote:

Two computers running Win11 Pro.

One has RTX4060,

One has RTX4080 Super.

Can I run these headless as is, or do I need to run spoof plugs? I don't know the correct term for it.

This is all for remote access to save electricity and space. Mostly space. My tables are full.

Phil

How much WUs simultaneously are you running on 4060? Asking for my newest config (work). 

Phil
Phil
Joined: 8 Jun 14
Posts: 849
Credit: 677309313
RAC: 9622319

KLiK wrote:Phil wrote:Two

KLiK wrote:

Phil wrote:

Two computers running Win11 Pro.

One has RTX4060,

One has RTX4080 Super.

Can I run these headless as is, or do I need to run spoof plugs? I don't know the correct term for it.

This is all for remote access to save electricity and space. Mostly space. My tables are full.

Phil

How much WUs simultaneously are you running on 4060? Asking for my newest config (work). 

all my 4060s are running x2 Performance Level 3 with temps hovering 60-65C.

Edit: Also, all machines with Nvidia gpus have been migrated to Ubuntu 24.04

 

 

Phil

I thought I was wrong once, but I was mistaken.

stfn
stfn
Joined: 7 Jun 21
Posts: 26
Credit: 95942852
RAC: 54791

Today I switched my ASUS 2060

Today I switched my ASUS 2060 6GB to an EVGA 3060 12GB, mostly for the VRAM.

The EVGA one is smaller in size, but heavier, the heatsink is a larger part of the whole package. What is great is that in the 3060 I can reduce the power ceiling of the GPU to 100W in nvidia-smi, I could only go as low as 130W in the 2060. What is not great is that nvtop is clearly showing that my GPU is now being bottlenecked by my old CPU. Probably this is the reason why I am only seeing a difference of a few percent in runtimes in O3AS.

KLiK
KLiK
Joined: 1 Apr 14
Posts: 92
Credit: 517036365
RAC: 1136596

Phil wrote:KLiK

Phil wrote:

KLiK wrote:

Phil wrote:

Two computers running Win11 Pro.

One has RTX4060,

One has RTX4080 Super.

Can I run these headless as is, or do I need to run spoof plugs? I don't know the correct term for it.

This is all for remote access to save electricity and space. Mostly space. My tables are full.

Phil

How much WUs simultaneously are you running on 4060? Asking for my newest config (work). 

all my 4060s are running x2 Performance Level 3 with temps hovering 60-65C.

Edit: Also, all machines with Nvidia gpus have been migrated to Ubuntu 24.04

Thanks, as work 4060 does ~8min job on Asteroids@home at ~50°C.

While on Einstein@home on BRPs it does around ~11min jobs at ~55°C & ASGW job for around ~50min. 

 

What are your ASGW times for dual processing?

 

Phil
Phil
Joined: 8 Jun 14
Posts: 849
Credit: 677309313
RAC: 9622319

KLiK wrote:Phil

KLiK wrote:

Phil wrote:

KLiK wrote:

Phil wrote:

Two computers running Win11 Pro.

One has RTX4060,

One has RTX4080 Super.

Can I run these headless as is, or do I need to run spoof plugs? I don't know the correct term for it.

This is all for remote access to save electricity and space. Mostly space. My tables are full.

Phil

How much WUs simultaneously are you running on 4060? Asking for my newest config (work). 

all my 4060s are running x2 Performance Level 3 with temps hovering 60-65C.

Edit: Also, all machines with Nvidia gpus have been migrated to Ubuntu 24.04

Thanks, as work 4060 does ~8min job on Asteroids@home at ~50°C.

While on Einstein@home on BRPs it does around ~11min jobs at ~55°C & ASGW job for around ~50min. 

 

What are your ASGW times for dual processing?

 

03AS tasks run around 4600 seconds at X2 per card on my 4060s.

 

 

 

Phil

I thought I was wrong once, but I was mistaken.

Tom M
Tom M
Joined: 2 Feb 06
Posts: 6862
Credit: 9774258650
RAC: 2903753

The Titan V has certain

The Titan V has certain features like a huge memory bus with HBM. 

I am wondering what model was it's direct descent?

Anything?

What kind of performance gain?

A Proud member of the O.F.A.  (Old Farts Association).

Ian&Steve C.
Ian&Steve C.
Joined: 19 Jan 20
Posts: 4152
Credit: 49826021594
RAC: 39950834

Tom M wrote: The Titan V has

Tom M wrote:

The Titan V has certain features like a huge memory bus with HBM. 

I am wondering what model was it's direct descent?

Anything?

What kind of performance gain?

the Turing line of cards came after the Titan V, but not exactly a successor other than the inclusion of Tensor cores. 

the V100 is the fully enabled version of the Titan V. Basically just more memory and more memory bandwidth. 

the A100 is probably the true successor. The architecture is the most similar. 

_________________________________________________________________________

KLiK
KLiK
Joined: 1 Apr 14
Posts: 92
Credit: 517036365
RAC: 1136596

Phil wrote:KLiK

Phil wrote:

KLiK wrote:

Phil wrote:

KLiK wrote:

Phil wrote:

Two computers running Win11 Pro.

One has RTX4060,

One has RTX4080 Super.

Can I run these headless as is, or do I need to run spoof plugs? I don't know the correct term for it.

This is all for remote access to save electricity and space. Mostly space. My tables are full.

Phil

How much WUs simultaneously are you running on 4060? Asking for my newest config (work). 

all my 4060s are running x2 Performance Level 3 with temps hovering 60-65C.

Edit: Also, all machines with Nvidia gpus have been migrated to Ubuntu 24.04

Thanks, as work 4060 does ~8min job on Asteroids@home at ~50°C.

While on Einstein@home on BRPs it does around ~11min jobs at ~55°C & ASGW job for around ~50min. 

 

What are your ASGW times for dual processing?

03AS tasks run around 4600 seconds at X2 per card on my 4060s.

You can run x2 per card, but there is no advantage with new v1,16 app...as the times are (more than) doubled on 4060! ;)

 

Also, check out my thread here on Asteroids@home for speed of nVidia cards in their algorithm: https://asteroidsathome.net/boinc/forum_thread.php?id=1085

Phil
Phil
Joined: 8 Jun 14
Posts: 849
Credit: 677309313
RAC: 9622319

Everything is on hold until I

Everything is on hold until I get this Epyc up and running. Some of my other machines have problems which I have not had time to address. For example, one of them refuses to download extra work, only what is actively crunching.
When I figure out this whole Epyc thing with the Titan gpu I'll get back to work on the other ones.
Plus, I'll only be able to run one or two gpus on the Epyc machine because I'm out of power. Any more and I'll be above 80% loading on the breaker and that's not a good idea. The building manager told me they would install a couple of more circuits for me last month, but they still haven't.

 

Phil

I thought I was wrong once, but I was mistaken.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.