 |
|
Message boards : Number crunching : Hardware requirements for upcoming models
Message board moderation
Previous · 1 · 2 · 3 · 4 · 5
| Author | Message |
|---|---|
|
Send message Joined: 9 Oct 20 Posts: 690 Credit: 4,391,754 RAC: 6,918 |
Or quite commonly by the presence of a cathedral.In the 21st century we still go by that rubbish? And don't the politically correct folk accept a mosque instead? |
 Dave JacksonSend message Joined: 15 May 09 Posts: 4540 Credit: 19,039,635 RAC: 18,944 |
Processor speed is not that relevant here, what is though is available core count. IFS (and OpenIFS) is a highly parallel model. If I enable threading, there is a 2x speedup with 2 cores, 3.5x with 4 cores etc. I plan to release some tests on the dev site which will try out multi-core, higher model resolution (and hence memory) after the summer with assistance from the CPDN team. Today is the last day of meteorological summer. ;) |
|
Send message Joined: 9 Oct 20 Posts: 690 Credit: 4,391,754 RAC: 6,918 |
Today is the last day of meteorological summer. ;)The only true definition. Why do most people insist the longest day, June 21st, is the start of summer? That's entirely illogical. |
|
Send message Joined: 5 Aug 04 Posts: 1120 Credit: 17,202,915 RAC: 2,154 |
I wonder if Boinc has knowledge of your internet speed, and could send different types of task to different people? Which internet speed? They have this for each machine. Average upload rate 460.45 KB/sec Average download rate 29829.53 KB/sec Average turnaround time 2.47 days But they do not have this.. Speakeasy Speed Test Your Results History Timestamp Download Upload Latency Jitter Quality Score Test Server 8/31/2022 16:34:2 78.97 Mbps 89.19 Mbps 8 ms 0 ms Excellent stosat-plfi-01.sys.comcast.net.prod.hosts.ooklaserver.net 8/31/2022 16:21:32 79.55 Mbps 89.79 Mbps 7 ms 1 ms Excellent speedgauge2.optonline.net.prod.hosts.ooklaserver.net 8/22/2022 14:47:11 81.07 Mbps 89.10 Mbps 7 ms 1 ms Excellent nyc.speedtest.clouvider.net.prod.hosts.ooklaserver.net Which might they want? 
|
|
Send message Joined: 15 Jan 06 Posts: 637 Credit: 26,751,529 RAC: 653 |
I wonder if Boinc has knowledge of your internet speed, and could send different types of task to different people? If not, there could always be a tickbox in preferences to accept enormous tasks. We have asked for the ability to select the type of tasks we want for a long time. It seems to be against their ideology, which is always the most difficult to argue against. But I for one won't attempt to mix enormous tasks with ordinary ones. They either select them based on what my machine can do, or allow me to select them. |
|
Send message Joined: 9 Oct 20 Posts: 690 Credit: 4,391,754 RAC: 6,918 |
Either would do, and the first is fine, it's how well your computer tends to communicate with project servers.I wonder if Boinc has knowledge of your internet speed, and could send different types of task to different people? If those are your results from the same computer, they're wrong. Or has Boinc confused KB/sec with Kb/sec? It wouldn't surprise me, it's bug-ridden crap. Your first results suggest you can get 240 Mbit. |
|
Send message Joined: 5 Aug 04 Posts: 1120 Credit: 17,202,915 RAC: 2,154 |
Either would do, and the first is fine, it's how well your computer tends to communicate with project servers. They ARE from the same machine. The first set of numbers are reported by the climateprediction.net if I ask it to describe my machine. I very much doubt the reported download speed. They should come out up to about 75 Megabits/second. So the reported download speed is manifestly wrong. Of course the servers may be slower than my connection, but not dramatocally faster. So 240 Megabits/sec is ridiculous. By comparison, Rosetta reports: Average upload rate 343.51 KB/sec Average download rate 17994.96 KB/sec Average turnaround time 2.91 days which also reports download rate too fast. The second set of numbers I believe. They measure the speeds to and from my machine to the servers in question. Two are in NYC, and one is somewhere in Northern New Jersey. I believe them as they are close to what I am paying for on Verizon FiOS.
|
|
Send message Joined: 9 Oct 20 Posts: 690 Credit: 4,391,754 RAC: 6,918 |
By comparison, Rosetta reports:Boinc has a problem with kb and kB. A schoolboy error. It's full of them. Just divide the nonsense they give you by 8. I have a real download speed of 32Mbit. CPDN reports this as 1507.35 KB/sec which equates to 11.8Mbit, which is feasable. Rosetta reports precisely the same! I'd say they were getting the numbers from the same source, but why are yours different? |
|
Send message Joined: 5 Aug 04 Posts: 1120 Credit: 17,202,915 RAC: 2,154 |
Boinc has a problem with kb and kB. A schoolboy error. It's full of them. Just divide the nonsense they give you by 8. If I look at Rosetta, it does not report the same numbers as does ClimatePrediction, but they are both wrong in similar ways. CPDN claims: Average upload rate 107.73 KB/sec Average download rate 17059.98 KB/sec Average turnaround time 2.47 days Rosetta claims Average upload rate 343.51 KB/sec Average download rate 17994.96 KB/sec Average turnaround time 2.91 days MilkyWay claims Average upload rate 132.27 KB/sec Average download rate 25680.89 KB/sec Average turnaround time 0.92 days Universe claims Average upload rate 109.65 KB/sec Average download rate 15789.07 KB/sec Average turnaround time 7.01 days Which are the nonsense and which are correct? I could believe those upload rates so I would multiply those numbers by 8 to get Kbits/sec, but the download rates are more likely to be Kbits/sec so they need to be divided by 1000, not 8. or maybe they should be divided by both; i.e., 8000. It seems to me there is no reason for these speeds to be the same. They have different histories (over what period are they averaged?) There servers are in different places, their hardware can be presumed different... It should not surprise you that the turnaround times are different, since their work units certainly take different times and my priorities are set differently for each of them. Here are the typical speeds I get with Speakeasy Speed Test. Now these servers are within less than 100 miles of me. I assume ClimatePrediction and Rosetta are further away from me and their servers may be slower. Your Results History Timestamp Download Upload Latency Jitter Quality Score Test Server 9/1/2022 7:13:46 77.75 Mbps 89.10 Mbps 7 ms 2 ms Excellent stosat-plfi-01.sys.comcast.net.prod.hosts.ooklaserver.net 8/31/2022 16:34:2 78.97 Mbps 89.19 Mbps 8 ms 0 ms Excellent stosat-plfi-01.sys.comcast.net.prod.hosts.ooklaserver.net 8/31/2022 16:21:32 79.55 Mbps 89.79 Mbps 7 ms 1 ms Excellent speedgauge2.optonline.net.prod.hosts.ooklaserver.net
|
|
Dave Jackson Send message Joined: 15 May 09 Posts: 4540 Credit: 19,039,635 RAC: 18,944 |
Here are the typical speeds I get with Speakeasy Speed Test. Now these servers are within less than 100 miles of me. I assume ClimatePrediction and Rosetta are further away from me and their servers may be slower. Downloads from CPDN are all from Oxford. Trickle ups also go to Oxford. Uploads can go to universities or data centres in lots of different locations around the world. The latest Windows ones going to Tasmania for instance. With my connection, it is only when there is a problem with a server that it is the limiting factor rather than my own connection but I get that for those with much faster connections that will not be the case. |
©2024 cpdn.org
