Just popping in real quick before lunch break is over so I haven't parsed through everything in this thread. I do have my CPU overclocked to 2.2 so that may be part of the reason for time discrepancy.
I wondered if clocks were the reason for the discrepancy. I checked mine thinking I was running at reduced clocks or something. I have heatsinks and fans on my SBC and they never get hot so didn't think that was the reason.
Ran vgencmd against my pi and it is doing the full speed of 1.4Ghz on all cores. So that is not the reason.
But if @Iggy you are running at 2.1Ghz, you aren't running the same cpu or you are doing LN2 or something.
I wondered if clocks were the reason for the discrepancy. I checked mine thinking I was running at reduced clocks or something. I have heatsinks and fans on my SBC and they never get hot so didn't think that was the reason.
Ran vgencmd against my pi and it is doing the full speed of 1.4Ghz on all cores. So that is not the reason.
But if @Iggy you are running at 2.1Ghz, you aren't running the same cpu or you are doing LN2 or something.
What ARM Cortex cpu do you have.
He has the same one I do but is overclocking it, he said it has auto heat control downclocking built in to it.
I wondered if clocks were the reason for the discrepancy. I checked mine thinking I was running at reduced clocks or something. I have heatsinks and fans on my SBC and they never get hot so didn't think that was the reason.
Ran vgencmd against my pi and it is doing the full speed of 1.4Ghz on all cores. So that is not the reason.
But if @Iggy you are running at 2.1Ghz, you aren't running the same cpu or you are doing LN2 or something.
What ARM Cortex cpu do you have.
He has the same one I do but is overclocking it, he said it has auto heat control downclocking built in to it.
Ok . . . . but what ARM Cortex cpu do you have then?
I'm running a Rpi 4 400. It comes stock with a 1.8 compared to a 4 model B's 1.5Ghz. Slightly different processor, don't know about any other differences besides the processor speed. They both use the Cortex-A72 though perhaps a different L2.
Well if I go by the title of this thread, I assume that Iggy has a Raspberry Pi 4 Model 400 which is that keyboard computer.
And all RPi 4 models use the ARM Cortex A72 SoC cpu. So that explains why he processes tasks so fast compared to my RPi 3 Model B+ using the older ARM Cortex A53 cpu.
But what is confusing is that BOINC identifies all of our RPi's as ARM ARMv7 Processor when in fact his RPi 4 is an ARMv8 processor.
That is what has confused me since I kept thinking we all had ARMv7 processors or RPi 3's of various models.
My Nvidia Jetson Nano SBC is identified as a ARMv8 processor and is based on a ARM Cortex A57 cpu.
[Edit] Just figured it out. The cpu will be identified as ARMv7 if run on a 32bit OS and ARMv8 if run on a 64bit OS.
Well if I go by the title of this thread, I assume that Iggy has a Raspberry Pi 4 Model 400 which is that keyboard computer.
And all RPi 4 models use the ARM Cortex A72 SoC cpu. So that explains why he processes tasks so fast compared to my RPi 3 Model B+ using the older ARM Cortex A53 cpu.
But what is confusing is that BOINC identifies all of our RPi's as ARM ARMv7 Processor when in fact his RPi 4 is an ARMv8 processor.
That is what has confused me since I kept thinking we all had ARMv7 processors or RPi 3's of various models.
My Nvidia Jetson Nano SBC is identified as a ARMv8 processor and is based on a ARM Cortex A57 cpu.
[Edit] Just figured it out. The cpu will be identified as ARMv7 if run on a 32bit OS and ARMv8 if run on a 64bit OS.
Ah, I didn't realize you had a 3. I didn't know that about the cpu info in your edit, thanks for passing that on.
Well if I go by the title of this thread, I assume that Iggy has a Raspberry Pi 4 Model 400 which is that keyboard computer.
And all RPi 4 models use the ARM Cortex A72 SoC cpu. So that explains why he processes tasks so fast compared to my RPi 3 Model B+ using the older ARM Cortex A53 cpu.
But what is confusing is that BOINC identifies all of our RPi's as ARM ARMv7 Processor when in fact his RPi 4 is an ARMv8 processor.
That is what has confused me since I kept thinking we all had ARMv7 processors or RPi 3's of various models.
My Nvidia Jetson Nano SBC is identified as a ARMv8 processor and is based on a ARM Cortex A57 cpu.
[Edit] Just figured it out. The cpu will be identified as ARMv7 if run on a 32bit OS and ARMv8 if run on a 64bit OS.
Ah, I didn't realize you had a 3. I didn't know that about the cpu info in your edit, thanks for passing that on.
And I have a B model one not the keyboard model 400 Iggy has
I've been having too many units coming through as invalidated so I stopped work and I may try again sometime in the future to see if there is an improvement. My WCG's units are been consistently valid from the beginning so I'm considering it my main project for now. With my original account that dates back to Feb. of '06, the thrill was the search for proof of gravity waves and as that was established a few years ago, I'm not going to dwell on this issue.
The number of invalids should drop fast. Tom removed the problem stripe 85 tasks he was trying to gain convergence on and the cause of all the invalids. He said once we started to get 10% invalid ratio he would abort them and he was good on his promise.
So now you should not get any more of the problem tasks except for the occasional resend. I have not receive any of the stripe85 tasks since he stopped them being distributed.
Now all my tasks validate as usual.
[Edit]
Confused. Thought I was on another project, MW in fact. Now the stripe 84 tasks are causing a lot of invalids. Hope that Tom removes those also.
Just popping in real quick
)
Just popping in real quick before lunch break is over so I haven't parsed through everything in this thread. I do have my CPU overclocked to 2.2 so that may be part of the reason for time discrepancy.
I wondered if clocks were the
)
I wondered if clocks were the reason for the discrepancy. I checked mine thinking I was running at reduced clocks or something. I have heatsinks and fans on my SBC and they never get hot so didn't think that was the reason.
Ran vgencmd against my pi and it is doing the full speed of 1.4Ghz on all cores. So that is not the reason.
But if @Iggy you are running at 2.1Ghz, you aren't running the same cpu or you are doing LN2 or something.
What ARM Cortex cpu do you have.
Keith Myers wrote: I
)
He has the same one I do but is overclocking it, he said it has auto heat control downclocking built in to it.
mikey wrote: Keith Myers
)
Ok . . . . but what ARM Cortex cpu do you have then?
I'm running a Rpi 4 400. It
)
I'm running a Rpi 4 400. It comes stock with a 1.8 compared to a 4 model B's 1.5Ghz. Slightly different processor, don't know about any other differences besides the processor speed. They both use the Cortex-A72 though perhaps a different L2.
https://www.cnx-software.com/2020/11/02/raspberry-pi-400-review-benchmarks-vs-raspberry-pi-4/
Well if I go by the title of
)
Well if I go by the title of this thread, I assume that Iggy has a Raspberry Pi 4 Model 400 which is that keyboard computer.
And all RPi 4 models use the ARM Cortex A72 SoC cpu. So that explains why he processes tasks so fast compared to my RPi 3 Model B+ using the older ARM Cortex A53 cpu.
But what is confusing is that BOINC identifies all of our RPi's as ARM ARMv7 Processor when in fact his RPi 4 is an ARMv8 processor.
That is what has confused me since I kept thinking we all had ARMv7 processors or RPi 3's of various models.
My Nvidia Jetson Nano SBC is identified as a ARMv8 processor and is based on a ARM Cortex A57 cpu.
[Edit] Just figured it out. The cpu will be identified as ARMv7 if run on a 32bit OS and ARMv8 if run on a 64bit OS.
Keith Myers wrote: Well if I
)
Ah, I didn't realize you had a 3. I didn't know that about the cpu info in your edit, thanks for passing that on.
Iggy wrote: Keith Myers
)
And I have a B model one not the keyboard model 400 Iggy has
I've been having too many
)
I've been having too many units coming through as invalidated so I stopped work and I may try again sometime in the future to see if there is an improvement. My WCG's units are been consistently valid from the beginning so I'm considering it my main project for now. With my original account that dates back to Feb. of '06, the thrill was the search for proof of gravity waves and as that was established a few years ago, I'm not going to dwell on this issue.
The number of invalids should
)
The number of invalids should drop fast. Tom removed the problem stripe 85 tasks he was trying to gain convergence on and the cause of all the invalids. He said once we started to get 10% invalid ratio he would abort them and he was good on his promise.
So now you should not get any more of the problem tasks except for the occasional resend. I have not receive any of the stripe85 tasks since he stopped them being distributed.
Now all my tasks validate as usual.
[Edit]
Confused. Thought I was on another project, MW in fact. Now the stripe 84 tasks are causing a lot of invalids. Hope that Tom removes those also.