estimated runtime change for TF

Message boards : News : estimated runtime change for TF

Author	Message
Rantanplan Send message Joined: 9 Nov 18 Posts: 10 Credit: 78,758,703 RAC: 32,969	Message 10931 - Posted: 9 Aug 2025, 13:13:54 UTC
	But sorry, on amd gpus i have a huge impact to the running times of each wu. On a half gpu the boinc manager says just 17 hours !!! to complete. There must be very strong failure.
	ID: 10931 · Rating: 0 · rate: / Reply Quote

rebirther Volunteer moderator Project administrator Project developer Project tester Project scientist Send message Joined: 2 Jan 13 Posts: 7963 Credit: 44,538,964 RAC: 5	Message 10932 - Posted: 9 Aug 2025, 15:58:59 UTC - in response to Message 10931. Last modified: 9 Aug 2025, 16:13:10 UTC
	What is the real runtime of a WU on your GPU? You can do a quick seconds test with 1. extract the app zip file in a separate folder outside BOINC 2. create a worktodo.txt with Factor=53A0E6FF58FC2C919A08F02F47C50000,438477947,75,76 inside 3. run mfakto.exe In the first line you can see an estimated total runtime. And be in mind AMD needs a full CPU core.
	ID: 10932 · Rating: 0 · rate: / Reply Quote

Rantanplan Send message Joined: 9 Nov 18 Posts: 10 Credit: 78,758,703 RAC: 32,969	Message 10933 - Posted: 9 Aug 2025, 17:12:30 UTC - in response to Message 10932. Last modified: 9 Aug 2025, 17:14:32 UTC
	got assignment: exp=438477947 bit_min=75 bit_max=76 (17.45 GHz-days) Starting trial factoring M438477947 from 2^75 to 2^76 (17.45 GHz-days) Using GPU kernel "cl_barrett32_76_gs_2" Date Time \| class Pct \| time ETA \| GHz-d/day Sieve Wait Aug 09 19:15 \| 52 1.4% \| 6.380 1h40m \| 246.18 81206 0.00% But i wont have time now to exactly check this.
	ID: 10933 · Rating: 0 · rate: / Reply Quote

rebirther Volunteer moderator Project administrator Project developer Project tester Project scientist Send message Joined: 2 Jan 13 Posts: 7963 Credit: 44,538,964 RAC: 5	Message 10934 - Posted: 9 Aug 2025, 17:26:35 UTC - in response to Message 10933. Last modified: 9 Aug 2025, 17:29:12 UTC
	got assignment: exp=438477947 bit_min=75 bit_max=76 (17.45 GHz-days) Starting trial factoring M438477947 from 2^75 to 2^76 (17.45 GHz-days) Using GPU kernel "cl_barrett32_76_gs_2" Date Time \| class Pct \| time ETA \| GHz-d/day Sieve Wait Aug 09 19:15 \| 52 1.4% \| 6.380 1h40m \| 246.18 81206 0.00% But i wont have time now to exactly check this. 1h40m, that is the correct runtime, there is no need for a full run, was only a test. still trying to give an exact estimated runtime for BOINC. The BOINC devs are involved. There are other projects with the same issue.
	ID: 10934 · Rating: 0 · rate: / Reply Quote

Rantanplan Send message Joined: 9 Nov 18 Posts: 10 Credit: 78,758,703 RAC: 32,969	Message 10935 - Posted: 9 Aug 2025, 17:48:04 UTC - in response to Message 10934.
	Anwendung TF 0.46 (opencl_ati_101) Name TF_75-76_312-443M_wu_1486 Status Aktiv erhalten 09.08.2025 19:48:04 Ablaufdatum 14.08.2025 19:44:51 Ressourcen 1 CPU + 0.5 AMD/ATI GPUs Geschätzter Berechnungsaufwand 800 GFLOPs Prozessorzeit 00:00:04 Prozessor-Zeit seit dem letzten Checkpoint 00:00:04 bisherige Laufzeit 00:02:40 Geschätzte verbleibende Zeit 17:31:59 Fortschritt 0,254% benötigter Arbeitsspeicher 159,25 MB Größe des Arbeitspakets 108,81 MB Verzeichnis slots/13 Prozess-Nr. 9156 Fortschrittsrate 5,760% pro Stunde Ausführbare Datei wrapper_26018_windows_x86_64.exe Application Name TF Plan Class opencl_ati_101
	ID: 10935 · Rating: 0 · rate: / Reply Quote

rebirther Volunteer moderator Project administrator Project developer Project tester Project scientist Send message Joined: 2 Jan 13 Posts: 7963 Credit: 44,538,964 RAC: 5	Message 10936 - Posted: 9 Aug 2025, 18:12:12 UTC - in response to Message 10935.
	The estimated time will be decreasing faster. BOINC with GPU still have isses with rsc_fpops_est while CPU is fine.
	ID: 10936 · Rating: 0 · rate: / Reply Quote

Rantanplan Send message Joined: 9 Nov 18 Posts: 10 Credit: 78,758,703 RAC: 32,969	Message 10937 - Posted: 9 Aug 2025, 18:30:46 UTC - in response to Message 10936. Last modified: 9 Aug 2025, 18:31:09 UTC
	4.2 % done, ETA 17hours 10 minutes. A half gpu (RX 560)
	ID: 10937 · Rating: 0 · rate: / Reply Quote

rebirther Volunteer moderator Project administrator Project developer Project tester Project scientist Send message Joined: 2 Jan 13 Posts: 7963 Credit: 44,538,964 RAC: 5	Message 10938 - Posted: 9 Aug 2025, 18:32:35 UTC - in response to Message 10937. Last modified: 9 Aug 2025, 18:45:25 UTC
	4.2 % done, ETA 17hours 10 minutes. A half gpu (RX 560) Nearly half speed of my RX5500XT (~56min) As the estimated runtime is 800 GFLOPs the formula is totally wrong. Your card has 2000GFLOPS 800/2000 = 0.4sek estimated runtime.
	ID: 10938 · Rating: 0 · rate: / Reply Quote

Rantanplan Send message Joined: 9 Nov 18 Posts: 10 Credit: 78,758,703 RAC: 32,969	Message 10939 - Posted: 9 Aug 2025, 18:42:05 UTC Last modified: 9 Aug 2025, 18:44:22 UTC
	At more than 4 hours crunshing, i quit. Anwendung TF 0.46 (opencl_ati_101) Name TF_75-76_312-443M_wu_1486 Status Aktiv erhalten 09.08.2025 19:48:04 Ablaufdatum 14.08.2025 19:44:51 Ressourcen 1 CPU + 0.5 AMD/ATI GPUs Geschätzter Berechnungsaufwand 800 GFLOPs Prozessorzeit 00:00:09 Prozessor-Zeit seit dem letzten Checkpoint 00:00:09 bisherige Laufzeit 00:56:53 Geschätzte verbleibende Zeit 17:05:08 Fortschritt 5,258% benötigter Arbeitsspeicher 159,30 MB Größe des Arbeitspakets 102,46 MB Verzeichnis slots/13 Prozess-Nr. 9156 Fortschrittsrate 5,400% pro Stunde Ausführbare Datei wrapper_26018_windows_x86_64.exe Application Name TF Plan Class opencl_ati_101
	ID: 10939 · Rating: 0 · rate: / Reply Quote

rebirther Volunteer moderator Project administrator Project developer Project tester Project scientist Send message Joined: 2 Jan 13 Posts: 7963 Credit: 44,538,964 RAC: 5	Message 10940 - Posted: 9 Aug 2025, 18:47:18 UTC - in response to Message 10939. Last modified: 9 Aug 2025, 19:34:56 UTC
	You could check the stderr.txt in slot folder for better percent status. I have posted this issue in BOINC mailing list, hope for a solution.
	ID: 10940 · Rating: 0 · rate: / Reply Quote

Rantanplan Send message Joined: 9 Nov 18 Posts: 10 Credit: 78,758,703 RAC: 32,969	Message 10941 - Posted: 9 Aug 2025, 19:43:44 UTC - in response to Message 10940.
	Almost 2h , 10% crunshed, i had it all, i quit.
	ID: 10941 · Rating: 0 · rate: / Reply Quote

rebirther Volunteer moderator Project administrator Project developer Project tester Project scientist Send message Joined: 2 Jan 13 Posts: 7963 Credit: 44,538,964 RAC: 5	Message 10942 - Posted: 9 Aug 2025, 19:48:09 UTC - in response to Message 10941.
	Almost 2h , 10% crunshed, i had it all, i quit. Don't believe BM, check stderr.txt
	ID: 10942 · Rating: 0 · rate: / Reply Quote

Rantanplan Send message Joined: 9 Nov 18 Posts: 10 Credit: 78,758,703 RAC: 32,969	Message 10947 - Posted: 13 Aug 2025, 10:09:03 UTC - in response to Message 10942.
	ok , the last wu i crunched took less than 2 hours, ok it will work. I am crunching again.Greetings Btw it shows to need over 20 hours right now to complete.
	ID: 10947 · Rating: 0 · rate: / Reply Quote

rebirther Volunteer moderator Project administrator Project developer Project tester Project scientist Send message Joined: 2 Jan 13 Posts: 7963 Credit: 44,538,964 RAC: 5	Message 10948 - Posted: 13 Aug 2025, 10:14:14 UTC - in response to Message 10947.
	ok , the last wu i crunched took less than 2 hours, ok it will work. I am crunching again.Greetings Btw it shows to need over 20 hours right now to complete. Good to know, still no answer yet for a solution. What we have used is not for GPU as posted from BOINC main dev.
	ID: 10948 · Rating: 0 · rate: / Reply Quote

Previous · 1 · 2
Post to thread

Message boards : News : estimated runtime change for TF