|
|
Post by AnthroHeart on Aug 17, 2021 16:12:25 GMT
I can't feel the Repeater any longer, most likely due to passing over to afterlife. My devotion to Anubis seems to be bridging life and death. I spoke about that to my mentor, and he said "It is but a matter of state". Maybe these words will help you somehow. Everything is a Quantum State. |
|
|
|
Post by AnthroHeart on Nov 3, 2021 9:28:12 GMT
I had to update the links for Google Colab because I no longer have that paid Google Drive I used to have.
You probably will need Google Colab Pro to run it. When I didn't have that, it just stopped running immediately.
With Colab Pro you can choose up to 24 GB of RAM.
2) File, Open Notebook, and choose gitHub from the top tabs. 4) Click the magnifying glass to search. 5) Click on the notebook. 6) Run each line (play button) at a time. 7) You may need to get Colab Pro if it isn't running as long as it should. 8) You may need to go to Runtime and choose GPU and High Ram.
It may be that so many people started using Repeater on Colab that they made it so you need Pro to do that now.
BEFORE you go Pro, try setting High-RAM and GPU. I may have just forgotten to do that.
I am trying the Pro+ which is like $50/month but it only lets me use 10GB of RAM. When I had Pro before it let me do up to 24 GB of RAM.
For some reason it's not letting me pick High RAM.
|
|
|
|
Post by reden on Nov 4, 2021 14:26:31 GMT
I had to update the links for Google Colab because I no longer have that paid Google Drive I used to have.
You probably will need Google Colab Pro to run it. When I didn't have that, it just stopped running immediately.
With Colab Pro you can choose up to 24 GB of RAM.
2) File, Open Notebook, and choose gitHub from the top tabs. 4) Click the magnifying glass to search. 5) Click on the notebook. 6) Run each line (play button) at a time. 7) You may need to get Colab Pro if it isn't running as long as it should. 8) You may need to go to Runtime and choose GPU and High Ram.
It may be that so many people started using Repeater on Colab that they made it so you need Pro to do that now.
BEFORE you go Pro, try setting High-RAM and GPU. I may have just forgotten to do that.
I am trying the Pro+ which is like $50/month but it only lets me use 10GB of RAM. When I had Pro before it let me do up to 24 GB of RAM.
For some reason it's not letting me pick High RAM.
"When I didn't have that, it just stopped running immediately." I found the solution to this some time ago. I wondered why it kept closing immediately. Turns out Colab silently upgraded its default CUDA version to 11.1, which is not compatible with the Repeater. It is made for 10.1 and maybe below. Thankfully there's still a separate CUDA 10.1 install within Colab, so you call the compiler command with the cuda compiler found in the 10.1 folder, then it works. I think it is in /usr/bin somewhere. I may have to boot up my Colab to check. Another noteworthy thing is that 2 or 3 days ago I exhausted my Colab running "invisible GPU credit" and was only offered the option to run without one. Yesterday, it was fixed. |
|
|
|
Post by AnthroHeart on Nov 4, 2021 15:39:22 GMT
|
|
|
|
Post by reden on Nov 4, 2021 15:43:48 GMT
You should give it a more distinctive name. Even though the repos are separate. Anyways, the GPU code is !/usr/local/cuda-10.1/bin/nvcc intention_repeater_max_cuda.cu -O3 -o intention_repeater_max_cuda |
|
|
|
Post by AnthroHeart on Nov 4, 2021 15:53:15 GMT
You should give it a more distinctive name. Even though the repos are separate. Anyways, the GPU code is !/usr/local/cuda-10.1/bin/nvcc intention_repeater_max_cuda.cu -O3 -o intention_repeater_max_cuda Since I am on Google Colab Pro+ it works fine for me using normal nvcc. I also get to use 45GB of RAM when not using GPU.
The titles are different: Repeater_MAX.ipynb vs Repeater_MAX_CUDA.ipynb
|
|
|
|
Post by AnthroHeart on Nov 4, 2021 17:48:05 GMT
Also, to prevent timeout, in FireFox you can follow these directions:
Open the inspector view by typing Ctrl+ Shift + i and then clicking on console tab at top.
Paste the below code snippet at bottom of console and hit enter
function ClickConnect(){
console.log("Working");
document.querySelector("#top-toolbar > colab-connect-button").shadowRoot.querySelector("#connect").click();
}
setInterval(ClickConnect,60000)
|
|
|
|
Post by AnthroHeart on Nov 13, 2021 21:44:47 GMT
I'm not sure if it was mentioned elsewhere but there is also a Google Colab Android App.
It's not the greatest, but it does the job.
I had to disable screen auto-rotate because people say that causes the app to reload itself.
So I can run colab wherever I am.
|
|
|
|
Post by reden on Feb 27, 2023 17:57:43 GMT
As of recently, Tesla T4 GPUs began getting distributed to users, even Free ones.
These GPUs were released in 2018, are made on a 12 nm process, and have 2560 CUDA cores.
This is unlike the previous P4 GPUs, which were released in 2016, made on a 16nm process, with the same amount of CUDA cores.
However, T4s do 65 TFLOPS of performance, whereas P4s only do 5.5.
T4s achieve 1.2-1.3 EHz on a very long Intention, and when it is somewhat shortened, they achieve +2 EHz, a far cry from the 40-80 PHz of the past.
|
|
|
|
Post by AnthroHeart on Feb 27, 2023 20:29:46 GMT
As of recently, Tesla T4 GPUs began getting distributed to users, even Free ones. These GPUs were released in 2018, are made on a 12 nm process, and have 2560 CUDA cores. This is unlike the previous P4 GPUs, which were released in 2016, made on a 16nm process, with the same amount of CUDA cores. However, T4s do 65 TFLOPS of performance, whereas P4s only do 5.5. T4s achieve 1.2-1.3 EHz on a very long Intention, and when it is somewhat shortened, they achieve +2 EHz, a far cry from the 40-80 PHz of the past. How long can you run a session for? |
|
|
|
Post by reden on Feb 27, 2023 23:26:23 GMT
As of recently, Tesla T4 GPUs began getting distributed to users, even Free ones. These GPUs were released in 2018, are made on a 12 nm process, and have 2560 CUDA cores. This is unlike the previous P4 GPUs, which were released in 2016, made on a 16nm process, with the same amount of CUDA cores. However, T4s do 65 TFLOPS of performance, whereas P4s only do 5.5. T4s achieve 1.2-1.3 EHz on a very long Intention, and when it is somewhat shortened, they achieve +2 EHz, a far cry from the 40-80 PHz of the past. How long can you run a session for? I have observed 1 hour. |
|
|
|
Post by AnthroHeart on Feb 28, 2023 10:27:29 GMT
How long can you run a session for? I have observed 1 hour. Is it more powerful than running normal computer for like 8-12 hours? |
|
|
|
Post by reden on Feb 28, 2023 17:43:47 GMT
Is it more powerful than running normal computer for like 8-12 hours? I haven't gauged it, but it may be. |
|
|
|
Post by AnthroHeart on Feb 28, 2023 18:10:13 GMT
I haven't used Colab since November. Running on my home laptop is enough for me.
I do energy work regardless on my own.
|
|
|
|
Post by reden on Mar 4, 2023 21:34:25 GMT
I just observed "I am calm." running at 4.341 EHz in Colab.
I felt it at once.
|
|
