(archive)MISFIT

[swl551]

If #9 puts GPU72 in the drivers seat , MISFIT will never need to be changed to comply/support/adhere to future decisions in the GPU72 space.


See screenshot for visual interpretation of #9.

[chalsall]

Originally Posted by swl551

See screenshot for visual interpretation of #9.

Very cool!!!

And I agree with you and Dub... When "Let GPU72 decide" is selected even the Pledge will be ignored. "What Makes Sense" will optionally lower the Pledge level when the UA is MISFIT, but will never raise it.

Also, please forgive me for being pedantic, but I'd suggest that "GzDz to fetch" should read "GHzDs to fetch".

Thanks for your work Scott.

[swl551]

Originally Posted by chalsall

Very cool!!!

And I agree with you and Dub... When "Let GPU72 decide" is selected even the Pledge will be ignored. "What Makes Sense" will optionally lower the Pledge level when the UA is MISFIT, but will never raise it.

Also, please forgive me for being pedantic, but I'd suggest that "GzDz to fetch" should read "GHzDs to fetch".

Thanks for your work Scott.

I am ready to release 2.4.6 (Let GPU72 Decide) whenever you are ready on your end. If you prefer a closed beta I'll grab the usual suspects.

Yes to "GHzDs" as the abbreviation.

[chalsall]

Originally Posted by swl551

I am ready to release 2.4.6 (Let GPU72 Decide) whenever you are ready on your end. If you prefer a closed beta I'll grab the usual suspects.

"Let's light this candle!" -- Alan Shepard

[swl551]

VERSION 2.4.6
1. Implemented GPU72 work fetching option #9 (Let GPU72 Decide) for LLTF and DCTF fetching. This feature puts all decisions, other than how much work to fetch, in the hands of GPU72.com. Thanks to all for input and vision on how to best optimize this feature.

Get from http://mersenneforum.org/misfit/


Enjoy.

Scott

[chalsall]

Originally Posted by swl551

1. Implemented GPU72 work fetching option #9 (Let GPU72 Decide) for LLTF and DCTF fetching. This feature puts all decisions, other than how much work to fetch, in the hands of GPU72.com. Thanks to all for input and vision on how to best optimize this feature.

Thanks Scott!

[Chuck]

Originally Posted by swl551

VERSION 2.4.6
1. Implemented GPU72 work fetching option #9 (Let GPU72 Decide) for LLTF and DCTF fetching. This feature puts all decisions, other than how much work to fetch, in the hands of GPU72.com. Thanks to all for input and vision on how to best optimize this feature.

OK; got it going. Will be a day before the next work fetch.

[flashjh]

Testing now.

[ixfd64]

Are you still releasing the source code? The latest versions seem to be binary-only.

[swl551]

Originally Posted by ixfd64

Are you still releasing the source code? The latest versions seem to be binary-only.

I only release source code for select utilities. MISFIT will remain closed source to the general public until:
1. I elect to release it before I die.
2. If die and the if_i_die promise is acted upon by someone. See misfit site for conditions of if_i_die.
3. Release of source code to certain individuals under a Non Disclosure Agreement is possible if the conditions merit such an event.

Why am I so protective?
1. MISFIT is my hobby. I don't want 90 programmers telling me what I should have/could have done for XYZ...

2. I HAVE SEEN how ruthless some contributors to this forum can be if someone makes a mistake, and I won't put myself in a position to be on the receiving end of that.

Scott

[chalsall]

Originally Posted by swl551

I only release source code for select utilities. MISFIT will remain closed source to the general public until:
1. I elect to release it before I die.
2. If die and the if_i_die promise is acted upon by someone. See misfit site for conditions of if_i_die.
3. Release of source code to certain individuals under a Non Disclosure Agreement is possible if the conditions merit such an event.

That's cool. You have demonstrated you are trustworthy enough that closed-source code from you executing on a user's machine is unlikely to be malware.

Originally Posted by swl551

Why am I so protective?
1. MISFIT is my hobby. I don't want 90 programmers telling me what I should have/could have done for XYZ...

2. I HAVE SEEN how ruthless some contributors to this forum can be if someone makes a mistake, and I won't put myself in a position to be on the receiving end of that.

Please understand that this is not personal. Many of us here are from scientific and/or engineering backgrounds. It's quite common, and frankly expected, to leave emotions behind and just deal with the empirical facts. If you've made a mistake, admit it. If someone else has made a mistake, point it out.

And then go out after work and enjoy a beer or two together. It's a bit like rugby for geeks....