Re: [sage-devel] RAM to build Sage

2014-05-09 Thread Volker Braun
We do have a arm machine in Oxford with more than enough ram (afair 4GB). Dima did some work on finishing the arm port. In any case, its not a ram issue. On Friday, May 9, 2014 1:41:44 PM UTC+2, mmarco wrote: > > Well, testing, for instance, if everything runs in an ARM machine can be > painf

Re: [sage-devel] RAM to build Sage

2014-05-09 Thread Jeroen Demeyer
On 2014-05-09 13:41, mmarco wrote: And if we want to deliver ARM binaries, we need to test everything there. This does not have to be the case. In fact, at some point we had Solaris and OS X PPC binaries even though not everything was working 100% on those platforms. If there are a few files

Re: [sage-devel] RAM to build Sage

2014-05-09 Thread mmarco
Well, testing, for instance, if everything runs in an ARM machine can be painful. And if we want to deliver ARM binaries, we need to test everything there. El viernes, 9 de mayo de 2014 13:04:20 UTC+2, Jeroen Demeyer escribió: > > On 2014-05-09 10:55, Julien Puydt wrote: > >> But personally, I

Re: [sage-devel] RAM to build Sage

2014-05-09 Thread Jeroen Demeyer
On 2014-05-09 10:55, Julien Puydt wrote: But personally, I don't consider RAM usage by doctests an issue. Perhaps because you don't use old or limited computers? No, because I think that doctest coverage is more important than memory usage of doctests. It is perfectly possible to *use* Sage on

Re: [sage-devel] RAM to build Sage

2014-05-09 Thread Julien Puydt
Hi, Le 06/05/2014 19:39, Jeroen Demeyer a écrit : On 2014-05-06 16:53, John Cremona wrote: In practice, it may be hard to gather the informatino on how much RAM tests take You could run them under "ulimit -v" and see when they start to fail. But personally, I don't consider RAM usage by docte

Re: [sage-devel] RAM to build Sage

2014-05-06 Thread Jeroen Demeyer
On 2014-05-06 16:53, John Cremona wrote: In practice, it may be hard to gather the informatino on how much RAM tests take You could run them under "ulimit -v" and see when they start to fail. But personally, I don't consider RAM usage by doctests an issue. Jeroen. -- You received this message

Re: [sage-devel] RAM to build Sage

2014-05-06 Thread leif
On Tuesday, May 6, 2014 4:55:54 PM UTC+2, Volker Braun wrote: > Just like the # optional: internet tests, anything that we don't actually > test will be broken. Every different type of # optional tag will lead to > fewer tests being done. > sage: foo() # optional - RAM bar Nice idea. -leif

Re: [sage-devel] RAM to build Sage

2014-05-06 Thread Volker Braun
Just like the # optional: internet tests, anything that we don't actually test will be broken. Every different type of # optional tag will lead to fewer tests being done. I -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this

Re: [sage-devel] RAM to build Sage

2014-05-06 Thread John Cremona
On 6 May 2014 15:43, David Roe wrote: >> Anyway, there are certainly (a lot of?) doctests that take quite some time, >> but don't use extraordinarily much memory, so I think John's idea isn't that >> bad. > I am flattered, since praise from leif is quite hard to earn ;) In practice, it may be ha

Re: [sage-devel] RAM to build Sage

2014-05-06 Thread David Roe
> Anyway, there are certainly (a lot of?) doctests that take quite some time, > but don't use extraordinarily much memory, so I think John's idea isn't that > bad. +1 > -leif > > >> >> On Tuesday, May 6, 2014 11:26:47 AM UTC+2, John Cremona wrote: >>> >>> If these tests are known, they could perh

Re: [sage-devel] RAM to build Sage

2014-05-06 Thread leif
On Tuesday, May 6, 2014 12:53:21 PM UTC+2, Volker Braun wrote: > Just "# long time" would be enough IMHO. It should be understood that long > doctests tend to consume more memory. > Which is another argument for *not* sorting doctests by previous runtime, i.e., *don't* run all of these first,

Re: [sage-devel] RAM to build Sage

2014-05-06 Thread Volker Braun
Just "# long time" would be enough IMHO. It should be understood that long doctests tend to consume more memory. On Tuesday, May 6, 2014 11:26:47 AM UTC+2, John Cremona wrote: > > On 6 May 2014 10:15, Jeroen Demeyer > > wrote: > > On 2014-05-06 10:59, Julien Puydt wrote: > >> > >> The ARM c

Re: [sage-devel] RAM to build Sage

2014-05-06 Thread Volker Braun
Just "# long time" would be enough IMHO. It should be understood that long doctests tend to consume more memory. On Tuesday, May 6, 2014 11:26:47 AM UTC+2, John Cremona wrote: > > On 6 May 2014 10:15, Jeroen Demeyer > > wrote: > > On 2014-05-06 10:59, Julien Puydt wrote: > >> > >> The ARM c

Re: [sage-devel] RAM to build Sage

2014-05-06 Thread John Cremona
On 6 May 2014 10:15, Jeroen Demeyer wrote: > On 2014-05-06 10:59, Julien Puydt wrote: >> >> The ARM chromebook I use has 2Gb RAM and some swap ; it can build sage, >> but some tests can't run. I complained about it last year... > > Some tests (in particular, some related to elliptic curves) *do* n

Re: [sage-devel] RAM to build Sage

2014-05-06 Thread Jeroen Demeyer
On 2014-05-06 10:59, Julien Puydt wrote: The ARM chromebook I use has 2Gb RAM and some swap ; it can build sage, but some tests can't run. I complained about it last year... Some tests (in particular, some related to elliptic curves) *do* need more than 2GB of RAM. This is a known fact and I don

Re: [sage-devel] RAM to build Sage

2014-05-06 Thread Julien Puydt
Hi, Le 06/05/2014 00:14, William Stein a écrit : Does anybody know how much RAM is required to build Sage these days. With MAKE not set, typing "make" with sage-6.2.rc2 on a VM with 4GB RAM results in failure (see below). Of course, I'm not using swap, and of course with swap you can do anyth

Re: [sage-devel] RAM to build Sage

2014-05-05 Thread Jeroen Demeyer
On 2014-05-06 00:14, William Stein wrote: Hi, Does anybody know how much RAM is required to build Sage these days. With MAKE not set, typing "make" with sage-6.2.rc2 on a VM with 4GB RAM results in failure (see below). Of course, I'm not using swap, and of course with swap you can do anything

[sage-devel] RAM to build Sage

2014-05-05 Thread William Stein
Hi, Does anybody know how much RAM is required to build Sage these days. With MAKE not set, typing "make" with sage-6.2.rc2 on a VM with 4GB RAM results in failure (see below). Of course, I'm not using swap, and of course with swap you can do anything if you're willing to wait. That said, it wo