Readiness of asmgcc for x86_64 linux?
Hi, I saw the PyPy Status Blog post mentioning that there is a working asmgcc for x86_64 linux. I wonder if you could clarify the status of it a bit further. The last thing Jason Creighton wrote on the subject that I can find, from Aug 13, was: "...the bottom line is that the main goal of my GSoC was accomplished: A working 64-bit PyPy JIT. Hopefully I'll be able to complete asmgcc-64, and make the JIT even faster..." But the new Status Blog post says " It not only includes working 64bit JIT (merged into PyPy trunk), but also a working asmgcc for x86_64 linux platform, that makes it possible to run the JIT on this architecture with our advanced garbage collectors" So it sounds like he (or someone) DID get the Linux version of it working. Has it been merged into the trunk? Does it seem stable? You say: "Expect this to be a major selling point for the next PyPy release :-)" Do you have an estimate of when that'll come out? I'm looking forward to testing PyPy for some of our music recommendation code. The main thing holding me back so far is the lack of 64-bit support. The other thing in the way is that I need to use multiple cores. I can home-grow a solution for my needs, but it would be great if the python multiprocessing library were to be supported. I see "r77223 - in pypy/branch/fast-forward/pypy/module/_multiprocessing: . test" in the svn commit log, dated Tuesday of this week (http://permalink.gmane.org/gmane.comp.python.pypy.cvs/29865)... I'm hoping that means it's going to be supported soon? That would be really great. Thanks, Gary -- Gary Robinson CTO Emergent Discovery, LLC personal email: garyrob@me.com work email: grobinson@emergentdiscovery.com Company: http://www.emergentdiscovery.com Blog: http://www.garyrobinson.net
On Thu, Sep 23, 2010 at 11:51 AM, Gary Robinson <garyrob@me.com> wrote:
Yes, 64-bit support for asmgcc as merged, however there appears to be a performance issue with it, it's not nearly as fast as it should be. multiproccessing was added to the stdlib in 2.6, we have a fast-forward branch that's aiming to implement 2.7, so when it's released it will contain a multiprocessing module. Alex -- "I disapprove of what you say, but I will defend to the death your right to say it." -- Voltaire "The people's good is the highest law." -- Cicero "Code can always be simpler than you think, but never as simple as you want" -- Me
Thanks for your response Alex... I have a couple follow-up questions:
Yes, 64-bit support for asmgcc as merged, however there appears to be a performance issue with it, it's not nearly as fast as it should be.
Is this a matter that is getting PyPy developer attention, or is expected to in the relatively near future?
That's great news. Is there any estimate of when a fairly stable beta will be available? Thanks! Gary -- Gary Robinson CTO Emergent Discovery, LLC personal email: garyrob@me.com work email: grobinson@emergentdiscovery.com Company: http://www.emergentdiscovery.com Blog: http://www.garyrobinson.net On Sep 23, 2010, at 12:54 PM, Alex Gaynor wrote:
On Thu, Sep 23, 2010 at 5:28 PM, Gary Robinson <garyrob@me.com> wrote:
We're aware of it, and it will definitely happen before we do any sort of release.
Amaury or Benjamin could better say.
Alex -- "I disapprove of what you say, but I will defend to the death your right to say it." -- Voltaire "The people's good is the highest law." -- Cicero "Code can always be simpler than you think, but never as simple as you want" -- Me
-- Gary Robinson CTO Emergent Discovery, LLC personal email: garyrob@me.com work email: grobinson@emergentdiscovery.com Company: http://www.emergentdiscovery.com Blog: http://www.garyrobinson.net
"Longer than you want"
Ain't that the truth! :) And yet, if you felt like one of the following blocks if time was more likely than the others, I'd be very interested in knowing: < 6 months < 1 year < 2 years Or if you'd rather not say anything even at that level, I understand. Thanks! Gary On Sep 23, 2010, at 5:55 PM, Benjamin Peterson wrote:
0 to 1/3 is big progress. Sounds to me like the <6 mo time frame might be the right one. I unfortunately can't contribute code/patches at this point (I don't have the time-freedom or competence in JIT's or C++ to try), but when you have 64-bit code that supports the multiprocessing module ready for end-user testing, I'll be eager to install it and try it against my code, and of course I'll let you know of any problems. Sounds like it's not quite there yet, but I'll continue to follow developments. Perhaps popping up here sometimes to ask again about progress. I'm excited about the work on 64-bit and multiprocessing... Best, Gary -- Gary Robinson CTO Emergent Discovery, LLC personal email: garyrob@me.com work email: grobinson@emergentdiscovery.com Company: http://www.emergentdiscovery.com Blog: http://www.garyrobinson.net On Sep 24, 2010, at 12:56 PM, Amaury Forgeot d'Arc wrote:
On Fri, Sep 24, 2010 at 2:16 PM, Gary Robinson <garyrob@me.com> wrote:
I unfortunately can't contribute code/patches at this point (I don't have the time-freedom or competence in JIT's or C++ to try),
Just to be clear, the pypy python interpreter is written in RPython, and for most of the stuff that is missing on the fast-forward branch you don't need to know nothing about how the JIT work. Its a pain to learn RPython, but just to be clear on what you need to know. Are there any tasks that you can do with pure python? -- Leonardo Santagada
I would love to do it if I had the time-freedom. The information you give above tells me that I probably do have the skills -- I've written a ton of python code over the last 12 years or so. But I'm working to the point that I'm depriving my family already, because my company needs my full attention now. There will come a time when I'll be able to do it, and I suspect PyPy will still be able to benefit from code and patch contributors then. But, of course, that's easy to say. The point is that PyPy would can use more contributors right now, and I, unfortunately, can't be among them. But I will be an active tester when it's ready for end-user testing. -- Gary Robinson CTO Emergent Discovery, LLC personal email: garyrob@me.com work email: grobinson@emergentdiscovery.com Company: http://www.emergentdiscovery.com Blog: http://www.garyrobinson.net On Sep 24, 2010, at 2:29 PM, Leonardo Santagada wrote:
On Thu, Sep 23, 2010 at 11:51 AM, Gary Robinson <garyrob@me.com> wrote:
Yes, 64-bit support for asmgcc as merged, however there appears to be a performance issue with it, it's not nearly as fast as it should be. multiproccessing was added to the stdlib in 2.6, we have a fast-forward branch that's aiming to implement 2.7, so when it's released it will contain a multiprocessing module. Alex -- "I disapprove of what you say, but I will defend to the death your right to say it." -- Voltaire "The people's good is the highest law." -- Cicero "Code can always be simpler than you think, but never as simple as you want" -- Me
Thanks for your response Alex... I have a couple follow-up questions:
Yes, 64-bit support for asmgcc as merged, however there appears to be a performance issue with it, it's not nearly as fast as it should be.
Is this a matter that is getting PyPy developer attention, or is expected to in the relatively near future?
That's great news. Is there any estimate of when a fairly stable beta will be available? Thanks! Gary -- Gary Robinson CTO Emergent Discovery, LLC personal email: garyrob@me.com work email: grobinson@emergentdiscovery.com Company: http://www.emergentdiscovery.com Blog: http://www.garyrobinson.net On Sep 23, 2010, at 12:54 PM, Alex Gaynor wrote:
On Thu, Sep 23, 2010 at 5:28 PM, Gary Robinson <garyrob@me.com> wrote:
We're aware of it, and it will definitely happen before we do any sort of release.
Amaury or Benjamin could better say.
Alex -- "I disapprove of what you say, but I will defend to the death your right to say it." -- Voltaire "The people's good is the highest law." -- Cicero "Code can always be simpler than you think, but never as simple as you want" -- Me
-- Gary Robinson CTO Emergent Discovery, LLC personal email: garyrob@me.com work email: grobinson@emergentdiscovery.com Company: http://www.emergentdiscovery.com Blog: http://www.garyrobinson.net
"Longer than you want"
Ain't that the truth! :) And yet, if you felt like one of the following blocks if time was more likely than the others, I'd be very interested in knowing: < 6 months < 1 year < 2 years Or if you'd rather not say anything even at that level, I understand. Thanks! Gary On Sep 23, 2010, at 5:55 PM, Benjamin Peterson wrote:
0 to 1/3 is big progress. Sounds to me like the <6 mo time frame might be the right one. I unfortunately can't contribute code/patches at this point (I don't have the time-freedom or competence in JIT's or C++ to try), but when you have 64-bit code that supports the multiprocessing module ready for end-user testing, I'll be eager to install it and try it against my code, and of course I'll let you know of any problems. Sounds like it's not quite there yet, but I'll continue to follow developments. Perhaps popping up here sometimes to ask again about progress. I'm excited about the work on 64-bit and multiprocessing... Best, Gary -- Gary Robinson CTO Emergent Discovery, LLC personal email: garyrob@me.com work email: grobinson@emergentdiscovery.com Company: http://www.emergentdiscovery.com Blog: http://www.garyrobinson.net On Sep 24, 2010, at 12:56 PM, Amaury Forgeot d'Arc wrote:
On Fri, Sep 24, 2010 at 2:16 PM, Gary Robinson <garyrob@me.com> wrote:
I unfortunately can't contribute code/patches at this point (I don't have the time-freedom or competence in JIT's or C++ to try),
Just to be clear, the pypy python interpreter is written in RPython, and for most of the stuff that is missing on the fast-forward branch you don't need to know nothing about how the JIT work. Its a pain to learn RPython, but just to be clear on what you need to know. Are there any tasks that you can do with pure python? -- Leonardo Santagada
I would love to do it if I had the time-freedom. The information you give above tells me that I probably do have the skills -- I've written a ton of python code over the last 12 years or so. But I'm working to the point that I'm depriving my family already, because my company needs my full attention now. There will come a time when I'll be able to do it, and I suspect PyPy will still be able to benefit from code and patch contributors then. But, of course, that's easy to say. The point is that PyPy would can use more contributors right now, and I, unfortunately, can't be among them. But I will be an active tester when it's ready for end-user testing. -- Gary Robinson CTO Emergent Discovery, LLC personal email: garyrob@me.com work email: grobinson@emergentdiscovery.com Company: http://www.emergentdiscovery.com Blog: http://www.garyrobinson.net On Sep 24, 2010, at 2:29 PM, Leonardo Santagada wrote:
participants (5)
-
Alex Gaynor -
Amaury Forgeot d'Arc -
Benjamin Peterson -
Gary Robinson -
Leonardo Santagada