Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > comp.lang.c++ > #14642
| From | 88888 Dihedral <dihedral88888@googlemail.com> |
|---|---|
| Newsgroups | comp.lang.c++ |
| Subject | Re: HOW TO CUT FAT UNITS IN FAT LIBS NOT FIT IN L2? |
| Date | 2012-03-01 01:23 -0800 |
| Organization | http://groups.google.com |
| Message-ID | <28643765.382.1330593784269.JavaMail.geo-discussion-forums@pbfa4> (permalink) |
| References | <29840611.1498.1330449897212.JavaMail.geo-discussion-forums@pbux2> <cc1e3b27-88d5-454e-9d19-2248368277ce@z31g2000vbt.googlegroups.com> |
在 2012年2月29日星期三UTC+8上午3时38分39秒,Quint Rankid写道: > On Feb 28, 12:24 pm, 88888 Dihedral <dihedral88...@googlemail.com> > wrote: > > After checking some C++ libs, I think now the problem is not the expensive DRAM > > and the HD in tera bytes. > > > > It is the problem to let executables stayed in L2 in the same chip in nontrivial program executions to beat those hobbists and novices in the execution. > > > It's not clear to me what you want to accomplish. > > Do you wish to reduce the size of your executables? I've heard that > some compilers have an option for this. Maybe the linkers do too? > > Wouldn't whatever is in L2 cache get rolled out somewhere if another > program needs it? > > > > > I regard the cheap DRAM as the slow HD in the 90's. > > Maybe you should look into some sort of semiconductor based HD? Or is > that an oxymoron? > > Please clarify. 在 2012年2月29日星期三UTC+8上午3时38分39秒,Quint Rankid写道: > On Feb 28, 12:24 pm, 88888 Dihedral <dihedral88...@googlemail.com> > wrote: > > After checking some C++ libs, I think now the problem is not the expensive DRAM > > and the HD in tera bytes. > > > > It is the problem to let executables stayed in L2 in the same chip in nontrivial program executions to beat those hobbists and novices in the execution. > > > It's not clear to me what you want to accomplish. > > Do you wish to reduce the size of your executables? I've heard that > some compilers have an option for this. Maybe the linkers do too? > > Wouldn't whatever is in L2 cache get rolled out somewhere if another > program needs it? > > > > > I regard the cheap DRAM as the slow HD in the 90's. > > Maybe you should look into some sort of semiconductor based HD? Or is > that an oxymoron? > > Please clarify. OK, lets test programs do repeated random RW for some data in L1, and then L2, and then the dram, and then the HD first for the current mass produced available hardware. Then one can think about how to design programs with those parameters changed over the time.
Back to comp.lang.c++ | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread
HOW TO CUT FAT UNITS IN FAT LIBS NOT FIT IN L2? 88888 Dihedral <dihedral88888@googlemail.com> - 2012-02-28 09:24 -0800
Re: HOW TO CUT FAT UNITS IN FAT LIBS NOT FIT IN L2? Quint Rankid <qbr567@gmail.com> - 2012-02-28 11:38 -0800
Re: HOW TO CUT FAT UNITS IN FAT LIBS NOT FIT IN L2? 88888 Dihedral <dihedral88888@googlemail.com> - 2012-03-01 01:23 -0800
Re: HOW TO CUT FAT UNITS IN FAT LIBS NOT FIT IN L2? Juha Nieminen <nospam@thanks.invalid> - 2012-02-28 20:01 +0000
Re: HOW TO CUT FAT UNITS IN FAT LIBS NOT FIT IN L2? Dombo <dombo@disposable.invalid> - 2012-02-28 22:22 +0100
Re: HOW TO CUT FAT UNITS IN FAT LIBS NOT FIT IN L2? Ian Collins <ian-news@hotmail.com> - 2012-02-29 10:30 +1300
csiph-web