Don wrote:
We take the point of view that chess programming is already difficult enough without also trying to cover every possible combination of hardware and OS so we optimize for the most common non-entry level case. We even considered dropping 32 bit support but this is still going to take a long time to die off I'm afraid.
Me too!
The way I solve all these problems is simply:
* use GCC intrinsics
* change the GCC compile flags: either -msse2 or -msse4.2.
I do not have a single #ifdef or inline assembly in my code, and that way I achieve maximum portability with minimal effort. The most common "non entry level case" is obviously x86-64 SSE4.2 for which I know that the code uses hardware 'popcunt' and 'bsfq'. For all other cases, I just don't care, and trust GCC to do what is best.
You're right, the combination of possibilities is just insane, and there's no way to do a compile optimized for each target. And it's not just about Intel and AMD but also all the ARM stuff with Phones and Tablets now...
Theory and practice sometimes clash. And when that happens, theory loses. Every single time.
Don wrote:
The next release is going to be a bug fix release first of all. The book no longer works in Komodo. In windows there is also a minor memory leak and it appears to be in the C++11 libraries as it does not happen on any other platform. The workaround if you are a tester is to start the engine between games, which all testers should do anyway. In Linux the timing does not work right on modern distributions such as Ubuntu 13.04 so I'm trying to fix that too. Should be within a couple of days.
Hi Don,
As in my setup (IPON) I do not use a book and the engine is restarted with every game my question is if the bug fix is strength relevant and the engine has to be retested?
Don wrote:
The next release is going to be a bug fix release first of all. The book no longer works in Komodo. In windows there is also a minor memory leak and it appears to be in the C++11 libraries as it does not happen on any other platform. The workaround if you are a tester is to start the engine between games, which all testers should do anyway. In Linux the timing does not work right on modern distributions such as Ubuntu 13.04 so I'm trying to fix that too. Should be within a couple of days.
Hi Don,
As in my setup (IPON) I do not use a book and the engine is restarted with every game my question is if the bug fix is strength relevant and the engine has to be retested?
Thanks
Ingo
The bug fix may add 1 or 2 ELO. From my point of view that is not worth testing again. Due the error margins of the test it could easily end up looking 3 or 4 ELO better or 3 or 4 ELO worse almost at random.
Don
Capital punishment would be more effective as a preventive measure if it were administered prior to the crime.
Don wrote:We even considered dropping 32 bit support but this is still going to take a long time to die off I'm afraid.
Indeed, it would be premature to drop 32 bit support. There is a huge (and growing) Android market and all Android devices are 32bit only.
I really meant 32 bit support for windows releases. Due to a non-compete agreement I cannot market Komodo on mobile devices until 2015 but I would like to support that later.
Capital punishment would be more effective as a preventive measure if it were administered prior to the crime.
Don wrote:
The bug fix may add 1 or 2 ELO. From my point of view that is not worth testing again. Due the error margins of the test it could easily end up looking 3 or 4 ELO better or 3 or 4 ELO worse almost at random.
Perfect, I will simply substitute Komodo 5.1 with Komodo 5.1a (or whatever the name might be).
Don wrote:
The bug fix may add 1 or 2 ELO. From my point of view that is not worth testing again. Due the error margins of the test it could easily end up looking 3 or 4 ELO better or 3 or 4 ELO worse almost at random.
Perfect, I will simply substitute Komodo 5.1 with Komodo 5.1a (or whatever the name might be).
Looking forward to Komodo 6
Ingo
I was thinking Komodo 5.1r2 probably where the r2 means release 2.
If you decide to re-test however you could wait for cooler weather - there would be no rush and no pressing need.
Capital punishment would be more effective as a preventive measure if it were administered prior to the crime.
Don wrote:
The bug fix may add 1 or 2 ELO. From my point of view that is not worth testing again. Due the error margins of the test it could easily end up looking 3 or 4 ELO better or 3 or 4 ELO worse almost at random.
Perfect, I will simply substitute Komodo 5.1 with Komodo 5.1a (or whatever the name might be).
Looking forward to Komodo 6
Ingo
I was thinking Komodo 5.1r2 probably where the r2 means release 2.
If you decide to re-test however you could wait for cooler weather - there would be no rush and no pressing need.
Right now we just have 15°C. It was 20° more last week!
Don wrote:
The bug fix may add 1 or 2 ELO. From my point of view that is not worth testing again. Due the error margins of the test it could easily end up looking 3 or 4 ELO better or 3 or 4 ELO worse almost at random.
Perfect, I will simply substitute Komodo 5.1 with Komodo 5.1a (or whatever the name might be).
Looking forward to Komodo 6
Ingo
I was thinking Komodo 5.1r2 probably where the r2 means release 2.
If you decide to re-test however you could wait for cooler weather - there would be no rush and no pressing need.
I just realize that, as 5.1r2 will be more than 10 Elo weaker than CCT, my single list will not test it. So for me it doesn't matter. That saves time and money
Don wrote:
The bug fix may add 1 or 2 ELO. From my point of view that is not worth testing again. Due the error margins of the test it could easily end up looking 3 or 4 ELO better or 3 or 4 ELO worse almost at random.
Perfect, I will simply substitute Komodo 5.1 with Komodo 5.1a (or whatever the name might be).
Looking forward to Komodo 6
Ingo
I was thinking Komodo 5.1r2 probably where the r2 means release 2.
If you decide to re-test however you could wait for cooler weather - there would be no rush and no pressing need.
I just realize that, as 5.1r2 will be more than 10 Elo weaker than CCT, my single list will not test it. So for me it doesn't matter. That saves time and money
Bye
Ingo
Yes, there is really no need.
Don
Capital punishment would be more effective as a preventive measure if it were administered prior to the crime.
