Posts by Jesse Viviano

41) Message boards : News : ATI application updated to 0.60 (Message 47873) Posted 15 Apr 2011 by Jesse Viviano Post: Actually, the plan class should be "ati14". The reason CAL 1.3 had plan classes "ati13", "ati13amd", and "ati13ati" is that AMD changed its libraries' names from AMD names to ATI names. AMD did not do any such thing for CAL 1.4. The "ati13ati" class refers to the ATI named libraries, and the others refer to the AMD named libraries.
42) Message boards : News : ATI application updated to 0.60 (Message 47817) Posted 14 Apr 2011 by Jesse Viviano Post: I am getting this message: 4/13/2011 11:26:17 PM Milkyway@home Message from server: ATI GPU: Upgrade to the latest driver to process tasks using your computer's GPU I do not know why I am getting this message because I am using AMD's latest stable driver package, Catalyst 11.3. The driver itself is version 8.831.2.0 for Windows 7 64-bit. Did the administrator set this to accept Catalyst 11.4 and later only? That package is still a beta package.
43) Message boards : Number crunching : Wrong application served to my computer (Message 47754) Posted 13 Apr 2011 by Jesse Viviano Post: One valid reason for keeping the 32-bit version is to take advantage of the x87 FPU's native 80-bit precision. However, using SSE2 throws that advantage out the window because SSE2 does things either in 32-bit or 64-bit precision and cannot handle 80-bit precision. The x87 FPU is deprecated in 64-bit mode by both AMD's ABI (used everywhere except in Windows) and by the Microsoft 64-bit ABI (used only in Windows), so it should not be used in 64-bit mode because neither ABI are required to save the contents of the x87 FPU registers when they need to swap tasks. They are required to save the contents of the x87 FPU registers for a program running in compatibility mode (e.g. a 16-bit or 32-bit protected mode program that makes absolutely no use of real or virtual 8086 mode running on an OS running in long mode).
44) Message boards : Number crunching : help!! lost several work units (Message 47704) Posted 12 Apr 2011 by Jesse Viviano Post: Since this project is not set to do resends, the only solution is to detach and reattach your computer. Then, MilkyWay@home will declare the tasks you took as invalid due to being detached, so you will be able to get new ones as well as freeing the lost ones for someone else to crunch.
45) Message boards : Number crunching : Wrong application served to my computer (Message 47606) Posted 11 Apr 2011 by Jesse Viviano Post: I noticed that my computer, a self-built computer with a Core i7 980X running Windows 7 64-bit, was served the 32-bit SSE2 application in an earlier CPU (not n-body) MilkyWay@home task. Later, it was served the non-SSE2 32-bit application in a later CPU reguler (not n-body) MilkyWay@home task. Since AMD64 and Intel's clone of that guarantee SSE2, why is the server serving me up the non-SSE2 application? Personally, I would prefer if the developers would develop a 64-bit multithreaded regular MilkyWay@home application. If it can be parallelized to run on a GPU, it should be easy to turn that into a multithreaded application.
46) Message boards : News : ATI application (v0.57) should be available now (Message 47497) Posted 10 Apr 2011 by Jesse Viviano Post: I have two errors I found last night: One of them is that if something suspends a task running on the ATI GPU client just as it begins, it will cause a computation error when the task is resumed. The other one is that if the ATI GPU client tries to chew on one of the de_separation_17_3s_fix_4_* series work units (which seem to error out on any client including CPU, Nvidia GPU, and ATI GPU clients), the GPU client sometimes fails to detect the computation error immediately and gets stuck at 0% forever until someone suspends and then resumes the task.
47) Message boards : Number crunching : Is the de_separation_17_3s_fix_4_* series broken? (Message 47495) Posted 10 Apr 2011 by Jesse Viviano Post: You need to update the driver on the HD 3850 to a driver that includes the APP driver. This provides the libraries needed to do GPGPU.
48) Message boards : Number crunching : Is the de_separation_17_3s_fix_4_* series broken? (Message 47382) Posted 10 Apr 2011 by Jesse Viviano Post: I finally found a work unit showing an Nvidia client choking on a work unit from this series along with an ATI client: work unit 1133. Now, we have work units from this series that fail on ATI/AMD GPUs (most of them from this series that I have seen), CPUs (work unit 999 which also includes ATI/AMD clients), and Nvidia GPUs (work unit 1133). Therefore, I think that this set of errors has nothing to do with the new GPU client. I also have noticed malloc errors and/or errors reading parameters in the stderr output of all of these work units in this series.
49) Message boards : News : ATI application (v0.57) should be available now (Message 47340) Posted 10 Apr 2011 by Jesse Viviano Post: I have been told to cross-post this error to this thread below because I am using stock applications. However, the task that got stuck has failed out by computation error, so I will update the post: I noticed a large number of work units in my work queue whose names start with "de_separation_17_3s_fix_4_", and they seem to generate plenty of compute errors. Could someone look into these work units? I have noticed that all CPU and ATI GPU clients used by my computer and others trying to handle this work have all returned computation errors. I cannot include Nvidia clients because none of these work units have been sent to an Nvidia client yet. Update: I am using Catalyst 11.3 on my reference Sapphire Radeon HD 6970. My OS is Windows 7 Ultimate 64-bit.
50) Message boards : Number crunching : Is the de_separation_17_3s_fix_4_* series broken? (Message 47338) Posted 10 Apr 2011 by Jesse Viviano Post: I am using stock applications, and the problems seem to be happening to all of the applications I have seen trying to chew on them and then choking on them (like former president Bush choking on a pretzel), including the CPU and the ATI GPU clients. Therefore, I doubt that this is related at all to the ATI GPU client if it is choking up CPU clients. Surprisingly, none of these that have landed in my queue have been sent to an Nvidia GPU yet. I will cross-post this in the thread you indicated, though.
51) Message boards : Number crunching : Is the de_separation_17_3s_fix_4_* series broken? (Message 47330) Posted 9 Apr 2011 by Jesse Viviano Post: I noticed a large number of work units in my work queue whose names start with "de_separation_17_3s_fix_4_", and they seem to generate plenty of compute errors. Could someone look into these work units? My GPU has spent twelve minutes on one of these and still is stuck at 0% progress, when it normally crunches a task in one to two minutes.
52) Message boards : News : sending out workunits (Message 47277) Posted 9 Apr 2011 by Jesse Viviano Post: I am getting download errors because the application to run these work units, "milkyway_0.57_windows_intelx86__ati14.exe", is not found on the server.
53) Message boards : News : reducing max_wus_in_progress (Message 47240) Posted 9 Apr 2011 by Jesse Viviano Post: The BOINC feeder is the program that maintains the pool of SQL database connections for the scheduler. It costs plenty of computing resources to set up a connection to an SQL database each time a scheduler request is made, so having this middleman saves these connections for reuse saves this overhead. The reason the scheduler cannot maintain this is that there really is no one "the scheduler". Each time the scheduler is called, a new scheduler process or thread (I am not sure which) is spawned and then is torn down at the end of the scheduler request. If the feeder is down, the scheduler cannot work.
54) Message boards : News : Old ATI applications deprecated (Message 47231) Posted 9 Apr 2011 by Jesse Viviano Post: BOINC is not yet ready for ATI/AMD OpenCL. You will have to wait until the 6.12.x series gets out of beta before BOINC officially supports OpenCL on ATI/AMD cards. PrimeGrid is using necessary trickery to get its ATI OpenCL applications running by disguising the application as a CAL 1.3 application, causing computation errors on HD 3xxx cards.
55) Message boards : Number crunching : No credit for validated result (Message 47073) Posted 6 Apr 2011 by Jesse Viviano Post: I noticed that I got zero credit when I had a work unit completed and validated. The zero credit work unit is work unit 267373660. This is for the MilkyWay@Home N-Body Simulation application and seems to not affect the regular MilkyWay@Home application.
56) Message boards : Number crunching : Aaargh! Servers are out of new work!(2)" (Message 46922) Posted 2 Apr 2011 by Jesse Viviano Post: From what I know about BOINC and the server status page, there is plenty of work units to process. However, the scheduler processes cannot get to them because they use the feeder to maintain a shared connection to the database via the feeder, which maintains one or more database connections for the scheduler processes to use because this saves the overhead of creating and destroying a connection to the database each time a new scheduler process is spawned to deal with each scheduler request. Since the feeder is down, there is no way for the scheduler to contact the database.
57) Message boards : Number crunching : Help with WU error (Message 46831) Posted 29 Mar 2011 by Jesse Viviano Post: Don't worry about these. The MilkyWay@home admins failed to cancel some old work units when they replaced the application that handles the MilkyWay@Home N-Body Simulation work units. These work units were built with the old engine in mind, and the new engine expects a different format. Just let them error out on your machine. Don't abort them because you will not be able to tell which ones are the old work units and which ones are designed for the new engine. At least the bad ones error out instantly.
58) Message boards : Number crunching : Validate errors (Message 46814) Posted 29 Mar 2011 by Jesse Viviano Post: Seeing that you have plenty of compute errors, I think that it might be your computer at fault. I think you are looking in the wrong area.
59) Message boards : Number crunching : Validate errors (Message 46812) Posted 29 Mar 2011 by Jesse Viviano Post: Here is one that had a validate error, and it was not a member of the de_separation_10_3s_free_1 run: de_separation_13_3s_free_1_65998_1301362073 also known as work unit 261694197.
60) Message boards : Number crunching : NVIDIA OpenCL result generation failing? (Message 46803) Posted 29 Mar 2011 by Jesse Viviano Post: While I am no fan of NVidia after one of its defective GPUs forced me to replace a motherboard in my laptop and therefore use an AMD GPU in my desktop, I noticed that many NVIDIA OpenCL results are failing recently. See 261630712 and 261632712 for examples of this. Work unit 261325920 shows two NVIDIA OpenCL work units that completed processing but needed verification, a third one error out, and then my result which confirmed the second NVIDIA OpenCL result and failed the first one. This suggests to me that something is screwing up the generation of NVIDIA OpenCL work units.

Previous 20 · Next 20