typo: "Hi, Javier..." innocuous but still...
On Sat, Apr 11, 2009 at 1:45 PM, Korey Sewell <[email protected]> wrote: > > However, different from SE in config.ini I can't just verify if is it >> running multithreaded since (I assume) the Linux scheduler will assign the >> workloads dynamically at runtime (with the scheduler). Is this what is >> happening, or would it simply run everything serially; no SMT? > > How, Javier, the short-answer is that it hasn't been tested for FS mode. > Thus, since although the O3 model has been coded for SMT execution there > aren't any linux-boot or FS regressions that make sure SMT works. I'm > assuming that the glue code that makes O3 work for 1 thread is pretty much > the same for 2,3, or even 4 threads but again regression testing is the key > word here. > > In fact, the SMT tests for O3_SE mode are hardly rigorous enough so that's > probably the 1st path to completion there. > > If you want to approximate SMT execution though, you can search the m5-dev > mail archives and recently there is a batch to run the SimpleCPU models in > MT mode for full system. Also, I believe there is an option to execute more > than one instruction per thread in the SimpleCPU. So if it makes sense to > you, you could adopt that model as an approximation. > > We'd love to have you help get O3_FS-SMT stable too, so if you want to take > that route please let us know. > > -Korey > > >> >> >> I wonder why would it not be supported if the Linux image is able to >> handle multiple CPU's why can't it handle multiple hard-threads, at least >> from 1 CPU? >> >> -- >> Javier >> >> On Sat, Apr 11, 2009 at 9:00 AM, <[email protected]> wrote: >> >>> Send m5-users mailing list submissions to >>> [email protected] >>> >>> To subscribe or unsubscribe via the World Wide Web, visit >>> http://m5sim.org/cgi-bin/mailman/listinfo/m5-users >>> or, via email, send a message with subject or body 'help' to >>> [email protected] >>> >>> You can reach the person managing the list at >>> [email protected] >>> >>> When replying, please edit your Subject line so it is more specific >>> than "Re: Contents of m5-users digest..." >>> >>> >>> Today's Topics: >>> >>> 1. Assertion `blkSize == pkt->getSize()' failed. (Devraj Chapagain) >>> 2. panic: Tried to execute unmapped address 0 (Devraj Chapagain) >>> 3. Re: O3 error in FS: Aborting before starting simulation >>> (Ali Saidi) >>> >>> >>> ---------------------------------------------------------------------- >>> >>> Message: 1 >>> Date: Fri, 10 Apr 2009 11:27:49 -0700 >>> From: Devraj Chapagain <[email protected]> >>> Subject: [m5-users] Assertion `blkSize == pkt->getSize()' failed. >>> To: M5 users mailing list <[email protected]> >>> Message-ID: >>> <[email protected]> >>> Content-Type: text/plain; charset="iso-8859-1" >>> >>> hi everyone,I am using 16 copies of SPEC CPU2006 benchmarks in SE mode >>> for >>> simulation. I got the error as: >>> >>> warn: Increasing stack size by one page. >>> m5.opt: build/ALPHA_SE/mem/cache/cache_impl.hh:312: bool >>> Cache<TagStore>::access(Packet*, typename TagStore::BlkType*&, int&, >>> PacketList&) [with TagStore = LRU]: Assertion `blkSize == pkt->getSize()' >>> failed. >>> >>> Could you please suggest me any solutions related to this problem. >>> >>> >>> Thanks in advance, >>> devraj >>> -------------- next part -------------- >>> An HTML attachment was scrubbed... >>> URL: >>> http://m5sim.org/cgi-bin/mailman/private/m5-users/attachments/20090410/bd0a177b/attachment.html >>> >>> ------------------------------ >>> >>> Message: 2 >>> Date: Fri, 10 Apr 2009 11:35:13 -0700 >>> From: Devraj Chapagain <[email protected]> >>> Subject: [m5-users] panic: Tried to execute unmapped address 0 >>> To: M5 users mailing list <[email protected]> >>> Message-ID: >>> <[email protected]> >>> Content-Type: text/plain; charset="iso-8859-1" >>> >>> hi everyone,While I am working on 8 core and 2 thread in SPECCPU 2006 >>> in SE mode, i am facing panic problem. >>> >>> hello robert white, >>> I saw you M5-users email correspondence with Ali, on this problem. On >>> that email you replied you solved this problem. Could you please help >>> me how to fix this problem, the problem is: >>> >>> warn: Increasing stack size by one page. >>> >>> panic: Tried to execute unmapped address 0. >>> >>> @ cycle 17190562000 >>> >>> [invoke:build/ALPHA_SE/arch/alpha/faults.cc, line 186] >>> >>> Program aborted at cycle 17190562000 >>> >>> >>> Thanks in advance, >>> >>> devraj >>> -------------- next part -------------- >>> An HTML attachment was scrubbed... >>> URL: >>> http://m5sim.org/cgi-bin/mailman/private/m5-users/attachments/20090410/40632096/attachment-0001.htm >>> >>> ------------------------------ >>> >>> Message: 3 >>> Date: Fri, 10 Apr 2009 17:23:33 -0500 >>> From: Ali Saidi <[email protected]> >>> Subject: Re: [m5-users] O3 error in FS: Aborting before starting >>> simulation >>> To: M5 users mailing list <[email protected]> >>> Message-ID: <[email protected]> >>> Content-Type: text/plain; charset=US-ASCII; format=flowed; delsp=yes >>> >>> >>> What kernel are you running? Have you modified the configuration in >>> any way? If you're using a one of our kernels and haven't modified the >>> configuration then the next thing to do would be to enable the trace >>> flags for the caches and the o3 cpu a few thousand cycles before the >>> panic occurs and see where the address in question is coming from. >>> >>> >>> Ali >>> >>> >>> >>> >>> On Apr 9, 2009, at 1:01 AM, Javier Jose wrote: >>> >>> > Hi everyone, >>> > >>> > I am trying to run ALPHA_FS with O3 but it is aborting before >>> > finishing regardless of the benchmark. >>> > >>> > COMMAND LINE: >>> > >>> > build/ALPHA_FS/m5.[any] -d /home/javier/log/mpgenc2_pvt configs/ >>> > example/fs.py -d --caches -b [benchmark] >>> > >>> > ERROR: >>> > >>> > info: Entering event queue @ 0. Starting simulation... >>> > panic: Unable to find destination for addr (user set default >>> > responder): 0x80c4dbc0 >>> > @ cycle 48692311000 >>> > [findPort:build/ALPHA_FS/mem/bus.cc, line 334] >>> > Memory Usage: 197392 KBytes >>> > For more information see: http://www.m5sim.org/panic/6c40f6ee >>> > Program aborted at cycle 48692311000 >>> > Aborted >>> > >>> > Right now I am just trying to run just 1 thread; 1 CPU. I have been >>> > using ALPBench and PARSEC with SimpleCPU and it works fine even with >>> > CMP configurations with L2 shared and L2 private cache configurations. >>> > >>> > Also, Is SMT still unsupported on FS? I can't seem to get a clear >>> > answer form the m5-user digest. >>> > >>> > I appreciate the help, >>> > >>> > Javier >>> > Texas A&M University >>> > Computer Science and Engineering Department >>> > _______________________________________________ >>> > m5-users mailing list >>> > [email protected] >>> > http://m5sim.org/cgi-bin/mailman/listinfo/m5-users >>> >>> >>> >>> ------------------------------ >>> >>> _______________________________________________ >>> m5-users mailing list >>> [email protected] >>> http://m5sim.org/cgi-bin/mailman/listinfo/m5-users >>> >>> End of m5-users Digest, Vol 33, Issue 14 >>> **************************************** >>> >> >> >> _______________________________________________ >> m5-users mailing list >> [email protected] >> http://m5sim.org/cgi-bin/mailman/listinfo/m5-users >> > > > > -- > ---------- > Korey L Sewell > Graduate Student - PhD Candidate > Computer Science & Engineering > University of Michigan > -- ---------- Korey L Sewell Graduate Student - PhD Candidate Computer Science & Engineering University of Michigan
_______________________________________________ m5-users mailing list [email protected] http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
