StarXTerminator 2.0.0 pre-release available for beta testing

[tom886]

Have you tried this solution?

Cheers
Tom

 

[johnbgood]

Thank you Tom and RCROMAN That worked. removed stars 45 seconds. That is Awesome.

Now I just need to get the other processes dialed in. Benchmark speed still very slow. I have 48 logical processors it should be ripping but only get Total 4116, CPU 3867, and SWAP 6174.

 

[aikivlado]

I have similar issue as posted in the past.Everything runs perfectly with 11 lite but Not with Full 11 version.I got the errors below:

*** Error:ERROR in ProcessTilel）. Error 8: 2 root error（s） found.
(0) RESOURCE_EXHAUSTED: DOM when allocating tensor with shape[1,32, 1024, 1024] and type float on /job: localhost/replica: 0/task:0/ device: GPU: 0 by allocator GPU_0_bfc
[LEnode starX/dd2_mconv/Conv2D11]]
Hint: If you want to see a list of allocated tensors when doM happens, add report_tensor_allocations_upon_com to RunOptions for current allocation info. This isn't available when running in Eager mode.
[[Identity_1/_5

Hint: If you want to see a list of allocated tensors when 00M happens, add report_tensor.
allocations_upon_oom to RunOptions for current
allocation info. This isn't available when running in Eager mode.


(1) RESOURCE_EXHAUSTED: OOM when allocating tensor with shape[1,32, 1024,1024] and type float on /job: localhost/replica:®/task:0/ device: GPU: 0 by allocator GPU_0_bf[LInode starx/dd2_mconv/Conv2D}}]]

Hint: If you want to see a list of allocated tensors when ooM happens, add report_tensor.
_allocations_upon_com to RunOptions for current
allocation info. This isn't available when running in Eager mode.


0 suCcessful operations.
0derived errors ignored.”””””


I am running on a gaming laptop with good characteristics:
13Gen Intel i9-13900H
Nvidia RTX 3050

Thank you!

 

[rcroman]

This is exactly why the "lite" version is supplied – it takes a lot of GPU memory to run the full version. Generally the recommendation is >6GB, but more could be required if other processes are using GPU RAM. More details here.

 

[aikivlado]

My appologies.I just saw that my GPU ram is just 4GB, which is big mistake from my side not paying attention when buying the laptop!
I read all the info from your post and thought my ram is way over that is why I asked the question. But now all make sense. Thank you for your quick reply.

 

[sn2006gy]

I got AMD GPU working with libtensorflow and wrote a blog here: https://rockchucksummit.com/pixinsight-gpu-acceleration-for-amd/ I'll try and download a demo of this script and verify this, but startnet++ worked for me in CLI and in PixInsight.

 

[sn2006gy]

Verified all the rc-astro tools worked. Downloaded a trial version of each and they're extremely fast on my 7900xtx. star mask was so fast on a 26mp image i didn't really even have time to switch windows to grab a screenshot of radeontop running on GPU.