I ran the memtester last night with a more realistic configuration last night and tripped on a bug some 4 hours later. I also had to increase the deadlock time in the memtester because the configuration results in larger delays.

system.cpu1: completed 1441000 read accesses @52151501002
system.cpu7: completed 1440000 read accesses @52155444000
system.cpu2: completed 1439000 read accesses @52156863003
system.cpu6: completed 1437000 read accesses @52157426000
system.cpu4: completed 1443000 read accesses @52160669006
system.cpu0: completed 1440000 read accesses @52168664002
system.cpu3: completed 1443000 read accesses @52171947004
system.cpu5: completed 1441000 read accesses @52186313001
system.cpu1: completed 1442000 read accesses @52190818002
system.cpu7: completed 1441000 read accesses @52191294004
system.cpu6: completed 1438000 read accesses @52192658003
system.cpu2: completed 1440000 read accesses @52195055002
system.cpu4: completed 1444000 read accesses @52195763000
panic: system.cpu7: read of 405507 (blk 405500) @ cycle 52199082001 returns f5, expected 55
 @ cycle 52199082001
[completeRequest:build/ALPHA_SE/cpu/memtest/memtest.cc, line 214]
Program aborted at cycle 52199082001
Aborted (core dumped)

Attachment: memtest.diff
Description: Binary data




Ali

On Mar 12, 2008, at 2:02 AM, Ali Saidi wrote:

Can you give us some information about what you're running configuration wise? Your config files would actually be good. You're stressing something that hasn't really been stressed before, so the first step is for us to be able to reliably duplicate the issue.

Ali




On Mar 12, 2008, at 1:27 AM, Sujay Phadke wrote:

I commented the assertion statement out. The benchmarks do complete, but calculations are incorrect for some of them. For example, radix outputs the following:

warn: ignoring syscall sigaction(5, 4831387552, ...)
warn: ignoring syscall sigaction(13, 4831387552, ...)
error with key 6475, value 102548 96863
error with key 7229, value 99786 97034
error with key 28571, value 178321 165003
error with key 28572, value 165003 141160
error with key 28573, value 141160 140297
error with key 51316, value 162386 146546
error with key 55007, value 165327 139551
error with key 55009, value 183538 165330

This didnt happen with beta 4.

- Sujay

----- Original Message ----- From: "Steve Reinhardt" <[EMAIL PROTECTED] >
To: "M5 users mailing list" <[EMAIL PROTECTED]>
Sent: Tuesday, March 11, 2008 5:38 PM
Subject: Re: [m5-users] error running splash2 in beta 5


That could be related to the cache changes I made to get writebacks to allocate in the cache. Looking at the assert, I'd say there's a 50/50
chance that the bug is simply that the assertion isn't valid anymore
rather than anything actually being broken.

Unfortunately I won't have time to look at it in detail for a few
days.  Can you try just commenting out the assertion and seeing if
those benchmarks complete successfully?

Thanks,

Steve

On Tue, Mar 11, 2008 at 2:08 PM, Sujay Phadke <[EMAIL PROTECTED] > wrote:


Hi all,
       I am trying to run the splash2 in m5 beta 5. On 5 of the
benchmarks, I get the following error:

m5.opt: build/ALPHA_SE/mem/cache/cache_impl.hh:1282: Packet*
Cache<TagStore>::getTimingPacket() [with TagStore = LRU]: Assertion
`tags->findBlock(mshr->addr) == __null' failed.

and M5 aborts. Can someone explain what this means?

- Sujay

_______________________________________________
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

_______________________________________________
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


_______________________________________________
m5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/m5-dev

Reply via email to