Posts by ahorek's team
|
21)
(Message 8912)
Posted 4 Apr 2025 by ahorek's team Post: Ray tracing improvements aren’t relevant for crunching. The new generation is expected to be slightly more efficient, but I think the 9070 XT and 7900 XT should deliver similar performance. Both cards should already be supported. Let us know once you have some results to share. |
|
22)
(Message 8911)
Posted 4 Apr 2025 by ahorek's team Post: Since it's an OpenCL app, the Radeon RX 9070 XT (RDNA 4) should work fine. However, for the card to be listed among the TOP GPUs, there must be at least 10 active hosts. This requirement ensures that the statistics remain meaningful. It looks like you don’t have the card and are just interested in its performance. Is that right? Unlike AMD, the official CUDA application doesn’t yet support Blackwell (GeForce RTX 5000 series), but the next version will. |
|
23)
(Message 8898)
Posted 1 Apr 2025 by ahorek's team Post: There's work available, and I'm receiving new tasks. What does the BOINC log say? Maybe wait a bit longer? |
|
24)
(Message 8890)
Posted 25 Mar 2025 by ahorek's team Post: Please refrain from insulting others. We're here to help, but we have no control over how the work is distributed. We're just volunteers like you. Some other projects, such as PrimeGrid, provide more consistent workloads. They can generate tasks that run for years with ease, making them a good choice if you want to keep your PC continuously occupied. On the other hand, GPUGrid has been out of work for a month. Are you suggesting they have "poor management" as well? |
|
25)
(Message 8878)
Posted 23 Mar 2025 by ahorek's team Post: Hi, the server recently experienced connection issues due to DDoS attacks. Additionally, the DNS changed this week, so if your local cache retained the old IP for a while, your PCs may have been unable to connect. Those issues are hopefully finally resolved. Unlike some other projects, Asteroids tasks require manual preprocessing and work loading. At times, work may not be available, this is normal and there's no need for concern. you can check the availability on the server status page https://asteroidsathome.net/boinc/server_status.php or https://grafana.kiska.pw/d/boinc/boinc?orgId=1&var-project=Asteroids@Home&from=now-1y&to=now |
|
26)
Posted 21 Mar 2025 by ahorek's team Post: Could you share the exact message or a screenshot of the warning you're getting? Perhaps you misinterpreted it. Avast does not identify the stock app as a threat. https://www.virustotal.com/gui/file/f49671fd6c0c9d293896b30a8ccc1b961c8fa1573ad90caf8155475d69899143/detection The app is not accessing any files unrelated to BOINC or .pdf files on its own. If your PC is already infected, some viruses may spread to any file they can access, including the Asteroids binary or other originally safe apps like your web browser, Windows internals, etc. |
|
27)
(Message 8862)
Posted 19 Mar 2025 by ahorek's team Post: Your CPU tasks are also failing/returning invalid results, even though they have been successfully validated by other users. This could be due to a hardware failure or an unstable overclock. You should check the stability of your NVidia system. how is a Bitcoin mining farm in Russia related to this project? :) |
|
28)
(Message 8859)
Posted 18 Mar 2025 by ahorek's team Post: Hi, a screensaver to visualize the current work was never finished. We have it in a roadmap, but it's not the current priority. I see that some of your NVidia GPU tasks crashed. Your driver should be recent enough, but updating it from 566.36 to 572.83 may help. Is the GPU overclocked? Do other GPU projects like Primegrid, Einstein, or games work reliably on that GPU? |
|
29)
Posted 15 Mar 2025 by ahorek's team Post: the university already has a good internet connection (1Gbps optics to the bone network) and resources to host a server. hosting the current server with the same performance and traffic (for instance on Amazon) would cost about $5000 per year The primary issue is a lack of human resources, with only one admin managing the server. A targeted attack can cause problems even for well-protected servers backed by an army of paid security technicians. This is just a small BOINC server, so the existing protections weren't as robust as those used by large companies running critical operations or businesses. It doesn't mean that no one cares. You know, it's easier to complain than fix the problem... the admin is already working on it, it just might take more time. |
|
30)
(Message 8856)
Posted 15 Mar 2025 by ahorek's team Post: Due to the DDoS attack, stricter access limits were enforced this week, to ensure the server remains accessible. Unfortunately, some regular BOINC users may also be affected by bans. This is unintentional, and the ban is only temporary. The admin is aware of the issue and is also working on a new firewall and servers to improve load handling and enhance protection against malicious actors. We all aim for a long-term resolution, but it may take some time. Since I don’t have access to the servers, I can't do anything to speed up the process. |
|
31)
(Message 8849)
Posted 12 Mar 2025 by ahorek's team Post: Apologies for the recent outages. Another DDoS attack hit the server, and it took the admin some time to secure it against those malicious actors. The web should be accessible now and we're still reviewing the configuration, but new work will be available soon. |
|
32)
(Message 8839)
Posted 28 Feb 2025 by ahorek's team Post: the released AMD app won't work unless it detects an AMD card. There are also differences in memory allocations, so the Intel app is slightly different, but we have an Intel build for testing. Battlemage is expected to work, unlike Alchemist, but it remains untested. Configuring the server to allow Intel devices and distributing the right app to them is another piece of the puzzle, but it shouldn't be that hard. |
|
33)
(Message 8835)
Posted 27 Feb 2025 by ahorek's team Post: Other projects typically don't require FP64. As long as they have an OpenCL app, they just need to enable Intel GPUs, and the existing application will work out of the box with no or minimal changes. There's no secret sauce they can share to help us support Intel GPUs. The Asteroids app heavily relies on FP64, but some GPUs, such as Intel ARC, Apple, or Adreno, don't support it at the hardware level. It's similar to trying to run an app designed specifically for AVX512 on a CPU that doesn't support it, it will simply crash. FP64 can be emulated, but it's not trivial, the app has to be rewritten and there's also a huge perf impact. Even if we manage them to work, I wouldn't expect a good performance. |
|
34)
(Message 8815)
Posted 21 Feb 2025 by ahorek's team Post: The primary goal of an NPU is power efficiency. It functions similarly to a video accelerator in a GPU, which excels at decoding video efficiently but it's unusable for anything else. You can decode a video on a general-purpose CPU or a GPU, but the efficiency won't be as good and it'll drain your battery faster. That’s also why desktop PCs don’t include an NPU, they don’t need one. NPUs also don't have a standardized API like an OpenCL Each vendor has their specific API, so it's hard to utilize them reliably on a platform like BOINC. New projects might find uses for an NPU in AI tasks, but it won't be technically usable for most current BOINC apps, even if developers wanted to. Remember that GPUs will continue to be significantly more powerful for AI and easier to work with, which is another reason I don’t expect any BOINC NPU projects in the near future. |
|
35)
(Message 8809)
Posted 19 Feb 2025 by ahorek's team Post: for Kepler cards, I suggest utilizing them on other OpenCL 1 based projects that may still work. Just keep in mind that the card is 12 years old and extremely power inefficient... |
|
36)
(Message 8808)
Posted 19 Feb 2025 by ahorek's team Post: > none of my Kepler cards are receiving tasks. latest drivers but specifically those two. sry, but Kepler cards don't support CUDA 5.5. https://www.techpowerup.com/gpu-specs/geforce-gtx-760.c1857 The CUDA 5.5 application is outdated (from 2016). It lacks support for certain tasks, causing failures like Number of lc points is greater than POINTS_MAX = 1000 We are working on a new version without these limitations and also plan to release the CUDA 5.5 version. However, those cards have been EOL for a long time and require a specialized application built on outdated software, which is becoming increasingly difficult to maintain. |
|
37)
(Message 8789)
Posted 12 Feb 2025 by ahorek's team Post: I tested the Intel version on the Core i3-6300, which supports DP, and it took around five hours. Since the iGPU usually slows down the CPU cores, it's unusable. https://www.techpowerup.com/gpu-specs/hd-graphics-530.c2789 Intel Arc Alchemist lacks FP64 support. I'm working on a software emulation to enable it, but it's still in the early stages. Even if all goes well, the performance will be limited. It’s more efficient to use the card for projects that don’t rely on FP64. Intel Arc Battlemage may work, but as far as I know, no one tested it. Volunteers willing to try are welcome :) |
|
38)
(Message 8785)
Posted 11 Feb 2025 by ahorek's team Post: you have to be a little bit more specific... Intel iGPUs (Intel HD Graphics) Intel Arc intel Battlemage? |
|
39)
(Message 8777)
Posted 7 Feb 2025 by ahorek's team Post: my Snapdragon 8 Gen 3 has 3 types of cores: 1 x 3.3 GHz ARM Cortex-X4 = the BIG one, fast 3 x 3.2 GHz ARM Cortex-A720 = cores for enhancing MT performance 2 x 2.3 GHz ARM Cortex-A520 = efficiency cores (useless for computing) Scaling is never perfect, note that all cores share the L3 cache and memory bandwidth. As a result, even without thermal constraints, utilizing all cores could potentially degrade performance. The same principle applies to Intel's P/E cores. Disabling E-cores is often beneficial, especially for multithreaded applications that share data. |
|
40)
(Message 8769)
Posted 4 Feb 2025 by ahorek's team Post: The FP32 marketing numbers don’t tell the whole story. Actual compute performance depends on how an application utilizes the GPU, each app may behave differently and memory bandwidth can also be a limiting factor. I recall when AMD had a specialized rotate instruction, while NVIDIA had to emulate it with four instructions to achieve the same result. In rare cases, such as SHA-256 calculations, AMD was significantly faster, but that feature was completely irrelevant for gaming. I agree that the situation has changed, and NVIDIA is now far ahead of the competition. While other vendors can still compete in gaming, NVIDIA typically holds a significant advantage when it comes to general computing performance. |
Previous 20 · Next 20