Could you please have a look at this post and the problem it points to and let us know if the idea of restricting the new releases to a smaller group of testers is workable for your purposes. In this case your latest release is restricted to SSE3 capable machines so hopefully that will be a smaller potential audience anyway.
Thanks, and thanks very much for your work.
I would like to use SSE3 in my optimized code, because it gives the best performance ( not much bigger, but bigger ). I know some people would like to get SSE2, SSE or only FPU optimized code, but i don't have enough time for each version. I spend at least 2 hours on a day to E@H ( coding, replies, etc. ) bacause i like it. But, what can i do? I spread my codes in 'patch' format and i help to Bernd to improve the performance of the official apps.
S5T0004 produces an invalid result on Pentium Mobile 740.
S5T0003 produces a valid result on my Pentium Mobile 740 and a valid hybrid result on AMD XP 3000+.
S5T0003 is around 2% faster than the default client.
Maybe the developers would also be interested in your patcher system/idea. Instead of boinc downloading an entire new app from the servers when theres an updated version it could just 'patch'. Just a thought as it could save a little bandwidth at times. Patches you make just a few Kb-whole app Mb or so..
Maybe the developers would also be interested in your patcher system/idea. Instead of boinc downloading an entire new app from the servers when theres an updated version it could just 'patch'. Just a thought as it could save a little bandwidth at times. Patches you make just a few Kb-whole app Mb or so..
I did the patcher system to prevent copyright things. I don't distribute any copyrighted code, and everybody can decide about its usage for his own risk.
Hm... I think a ZIP or an other file compressor could be help on the bandwith problem. Eg: My patcher program is also compressed and what is more this code is self-extractor so it consists a decompression program part too. You can check its size... Not so bad...
edit: but the code of patcher is very poor ( big ), i spent just some minutes to write it.
I would like to use SSE3 in my optimized code, because it gives the best performance ( not much bigger, but bigger ).
That sounds fine and I'm sure there would be a small enough group of people with SSE3 capable machines to test your latest patch files without risking the integrity of the validated results. Now that invalids have been seen, people are likely to be much more alert to the possibility in future.
Quote:
I know some people would like to get SSE2, SSE or only FPU optimized code, but i don't have enough time for each version.
There is a very large group of people who would have SSE or above who would be clamouring for a suitable patch. It was the largeness of this group that had me concerned. If you are not releasing further patches that would run on these lower spec machines then the integrity problem would largely disappear.
Quote:
I spend at least 2 hours on a day to E@H ( coding, replies, etc. ) bacause i like it. But, what can i do? I spread my codes in 'patch' format and i help to Bernd to improve the performance of the official apps.
You are doing a great job and using the patch format was a very good way of getting quick feedback without having to do the testing yourself. The community really appreciates your efforts. However, people can be very selfish and demanding so I expect that many will keep pestering you for more patches that will run on their particular hardware.
Here's a suggestion for you to consider so as to give yourself an income to justify the time you would need to spend. Set up a paypal account that people who really wanted an optimised app could donate to. Advise people of the details and invite them to donate whatever they felt an optimised app suitable to their cpu would be worth, with a minimum of say $USD10.00. Obviously people with bigger farms should donate more than a person with a single box. If insufficient donations were made you would not be obligated to do anything. Once there was sufficient you could produce patches and advise those who had donated of a location where the patches could be obtained. It would be an "honour" system that people who were told the location didn't reveal it to others.
Please realise that it is just an "off the top of the head" suggestion which would need to be carefully thought through. If it's too much trouble just forget about it. I only make the suggestion at all because I would like to see you get an ongoing reward for what you have already done and I firmly believe that there would be a large number of supporters here who would be willing to see you get an income from your efforts.
RE: When I do patching in
)
You got this message from the operation system not from the patcher. You should get a " Status: Done. " message if the patching is good.
RE: S5T0301 So, use this
)
How do I find out if I have SSE3?
RE: How do I find out if I
)
Download and run CPU-Z
www.boinc.cz
RE: RE: How do I find
)
thanx, that helped!
RE: @ Akos, Could you
)
I would like to use SSE3 in my optimized code, because it gives the best performance ( not much bigger, but bigger ). I know some people would like to get SSE2, SSE or only FPU optimized code, but i don't have enough time for each version. I spend at least 2 hours on a day to E@H ( coding, replies, etc. ) bacause i like it. But, what can i do? I spread my codes in 'patch' format and i help to Bernd to improve the performance of the official apps.
Thanks for the info, also
)
Thanks for the info, also thanks for your great work!
Welcome To Team China!
S5T0004 produces an invalid
)
S5T0004 produces an invalid result on Pentium Mobile 740.
S5T0003 produces a valid result on my Pentium Mobile 740 and a valid hybrid result on AMD XP 3000+.
S5T0003 is around 2% faster than the default client.
@ Akosf Maybe the
)
@ Akosf
Maybe the developers would also be interested in your patcher system/idea. Instead of boinc downloading an entire new app from the servers when theres an updated version it could just 'patch'. Just a thought as it could save a little bandwidth at times. Patches you make just a few Kb-whole app Mb or so..
RE: Maybe the developers
)
I did the patcher system to prevent copyright things. I don't distribute any copyrighted code, and everybody can decide about its usage for his own risk.
Hm... I think a ZIP or an other file compressor could be help on the bandwith problem. Eg: My patcher program is also compressed and what is more this code is self-extractor so it consists a decompression program part too. You can check its size... Not so bad...
edit: but the code of patcher is very poor ( big ), i spent just some minutes to write it.
RE: I would like to use
)
That sounds fine and I'm sure there would be a small enough group of people with SSE3 capable machines to test your latest patch files without risking the integrity of the validated results. Now that invalids have been seen, people are likely to be much more alert to the possibility in future.
There is a very large group of people who would have SSE or above who would be clamouring for a suitable patch. It was the largeness of this group that had me concerned. If you are not releasing further patches that would run on these lower spec machines then the integrity problem would largely disappear.
You are doing a great job and using the patch format was a very good way of getting quick feedback without having to do the testing yourself. The community really appreciates your efforts. However, people can be very selfish and demanding so I expect that many will keep pestering you for more patches that will run on their particular hardware.
Here's a suggestion for you to consider so as to give yourself an income to justify the time you would need to spend. Set up a paypal account that people who really wanted an optimised app could donate to. Advise people of the details and invite them to donate whatever they felt an optimised app suitable to their cpu would be worth, with a minimum of say $USD10.00. Obviously people with bigger farms should donate more than a person with a single box. If insufficient donations were made you would not be obligated to do anything. Once there was sufficient you could produce patches and advise those who had donated of a location where the patches could be obtained. It would be an "honour" system that people who were told the location didn't reveal it to others.
Please realise that it is just an "off the top of the head" suggestion which would need to be carefully thought through. If it's too much trouble just forget about it. I only make the suggestion at all because I would like to see you get an ongoing reward for what you have already done and I firmly believe that there would be a large number of supporters here who would be willing to see you get an income from your efforts.
Cheers,
Gary.