I just did an experiment with 2 of my 3 Raspberry Pi 4s, The two in question are running Buster/32-bit Debian.
Running 1 job takes ~3.6 hours. Or 6.67 jobs/day
Running 2 jobs takes ~4.5 hours Or 10.67 Jobs/day
Running 4 jobs takes ~5 hours. Or 19.20 jobs/day
So, my worry about cache misses causing slowness due to more cores accessing memory simultaneously was founded, but not by enough to stop running extra jobs. The numbers above are somewhat askew, as they were taken from too few samples, but I think are good enough that I am going to keep all 4 cores going on my 32Bit workaday Pis, and 2 jobs on my file server (which I did the experiments on, and just added to the mix yesterday). (The 1st row was only one sample.)
My 64-bit Pi, running a different (Beta still?), gets around 18.46 jobs/day on the same basis. Obviously not optimized for Bullseye. :(
you need to run the coolbits command again to enable overclocking for the new GPU. the command will run on all GPUs in the system at time of running it. so if you add a GPU later, you need to run it again to enable for the new GPU. I'm guessing when you first setup this system, you only had one GPU, then added another later on. or maybe you moved the GPU from one slot to another slot (changing the PCI address, and thus making it look like a new GPU).
Since it currently looks like my Nvidia 3 gpu rig is where I am going to be for a while, I just dropped the gpus to single tasks. By my WAG I will get up past 6M RAC again. And if/when it peaks, I can always "revert" to two tasks per GPU.
Now I can get the "feeling" of the processing speed of the Radeon VII I once had. Only 3 Radeon VII's equivalent :)
Tom M
A Proud member of the O.F.A. (Old Farts Association). Be well, do good work, and keep in touch.® (Garrison Keillor)
I have a singleton rtx 3080 here. By calculation it "should" be approaching 2.8 Million RAC on a "steady" diet of Einstein@Home's Gamma Ray #1 GPU project.
Instead it is peaking just over 1.8 Million RAC. So the question is, are my estimates that bad? Or is the strain the E@H servers are under making that kind of difference?
George?
Tom M
A Proud member of the O.F.A. (Old Farts Association). Be well, do good work, and keep in touch.® (Garrison Keillor)
there is a slump in validation due to the Pentathlon. many people holding tasks and not returning them until tomorrow. points are not awarded until tasks are validated. so if your wingman doesnt return the task yet, you wont get credit. expect a large spike in credit tomorrow.
but for the nth time, RAC is an average and not indicative of daily production. that system did 2.2 million yesterday. look at daily numbers, not RAC.
I have a singleton rtx 3080 here. By calculation it "should" be approaching 2.8 Million RAC on a "steady" diet of Einstein@Home's Gamma Ray #1 GPU project.
Instead it is peaking just over 1.8 Million RAC. So the question is, are my estimates that bad? Or is the strain the E@H servers are under making that kind of difference?
George?
Tom M
I'm here... I agree with Ian. Look at the daily numbers, not the average, for consistency.
The average will give you a performance outlook over a much longer period of time. It takes the highs and lows and 'averages' them out.
They’re gonna struggle to move it in this market. It’s barely faster than a 3080Ti/3090 (though will have better thermals than a 3090 since it has all mem modules on the front side of the PCB). Even with the “free” PSU, which isn’t really free, they’re just accounting for it in the price.
theyve been liquidating those 1600W (and others) PSUs with many sales on them for the last few months. I expect new PSUs with the new 12/16-pin connectors will be available in the near future.
Ronald McNichol wrote: I
)
When you set it up did you do this? https://pimylifeup.com/raspberry-pi-swap-file/
probably best to take RPi
)
probably best to take RPi discussions to the RPi thread in the stickies, since this thread is all about Nvidia.
_________________________________________________________________________
Zalster wrote: addiction is
)
. . Well of course, it's all about the science :)
. . How R U Zalster? Don't get to chat much since you know what closed down :)
Stephen
Ian&Steve C. wrote: you need
)
. . Especially that part :)
Stephen
Since it currently looks like
)
Since it currently looks like my Nvidia 3 gpu rig is where I am going to be for a while, I just dropped the gpus to single tasks. By my WAG I will get up past 6M RAC again. And if/when it peaks, I can always "revert" to two tasks per GPU.
Now I can get the "feeling" of the processing speed of the Radeon VII I once had. Only 3 Radeon VII's equivalent :)
Tom M
A Proud member of the O.F.A. (Old Farts Association). Be well, do good work, and keep in touch.® (Garrison Keillor)
I have a singleton rtx 3080
)
I have a singleton rtx 3080 here. By calculation it "should" be approaching 2.8 Million RAC on a "steady" diet of Einstein@Home's Gamma Ray #1 GPU project.
Instead it is peaking just over 1.8 Million RAC. So the question is, are my estimates that bad? Or is the strain the E@H servers are under making that kind of difference?
George?
Tom M
A Proud member of the O.F.A. (Old Farts Association). Be well, do good work, and keep in touch.® (Garrison Keillor)
there is a slump in
)
there is a slump in validation due to the Pentathlon. many people holding tasks and not returning them until tomorrow. points are not awarded until tasks are validated. so if your wingman doesnt return the task yet, you wont get credit. expect a large spike in credit tomorrow.
but for the nth time, RAC is an average and not indicative of daily production. that system did 2.2 million yesterday. look at daily numbers, not RAC.
_________________________________________________________________________
Tom M wrote: I have a
)
I'm here... I agree with Ian. Look at the daily numbers, not the average, for consistency.
The average will give you a performance outlook over a much longer period of time. It takes the highs and lows and 'averages' them out.
Proud member of the Old Farts Association
Here is a gpu + PSU
)
Here is a gpu + PSU announcement.
https://videocardz.com/newz/evga-rtx-3090-ti-kingpin-goes-on-sale-for-2500-usd-comes-with-free-1600w-psu">VideoCardz.com: EVGA RTX 3090 Ti KINGPIN goes on sale for 2500 USD, comes with 'free' 1600W PSU. https://videocardz.com/newz/evga-rtx-3090-ti-kingpin-goes-on-sale-for-2500-usd-comes-with-free-1600w-psu
A Proud member of the O.F.A. (Old Farts Association). Be well, do good work, and keep in touch.® (Garrison Keillor)
They’re gonna struggle to
)
They’re gonna struggle to move it in this market. It’s barely faster than a 3080Ti/3090 (though will have better thermals than a 3090 since it has all mem modules on the front side of the PCB). Even with the “free” PSU, which isn’t really free, they’re just accounting for it in the price.
theyve been liquidating those 1600W (and others) PSUs with many sales on them for the last few months. I expect new PSUs with the new 12/16-pin connectors will be available in the near future.
_________________________________________________________________________