Re: [m5-dev] Introducing...InOrder AlphaFS!

2011-04-21 Thread Gabe Black

Tada! Congratulations.

Gabe

On 4/21/2011 12:12 AM, Korey Sewell wrote:

For 1-4 wide, InOrder Cores running in ALPHA FS mode... Observe terminal
output:
M5 console: m5AlphaAccess @ 0xFD02
Got Configuration 623
memsize 800 pages 4000
First free page after ROM 0xFC018000
HWRPB 0xFC018000 l1pt 0xFC04 l2pt 0xFC042000
l3pt_rpb 0xFC044000 l3pt_kernel 0xFC048000 l2reserv
0xFC046000
kstart = 0xFC31, kend = 0xFC855898, kentry =
0xFC31, numCPUs = 0x1
CPU Clock at 2000 MHz IntrClockFrequency=1024
Booting with 1 processor(s)
...
Linux version 2.6.13 (h...@zed.eecs.umich.edu) (gcc version 3.4.3) #1 SMP
Sun Oct 8 ..
Brought up 1 CPUs
...
init started:  BusyBox v1.1.0 (2007.03.04-01:07+) multi-call binary
mounting filesystems...
EXT2-fs warning: checktime reached, running e2fsck is recommended
loading script...
*
**Two Words: Yes Sir!

*More Than Two Words: InOrder is booting Linux!!!

Getting to this point took about30 patches, so I have a lot of
patch-cleaning work to do before I can place it in the repo.

Also, just because it boots doesn't mean I've tested it with any kind of
real workload suite (e.g. SPEC or Parsec).  I don't have the bandwidth to
complete stress test through all the benchmarks and test but if someone
tries and something goes wrong, please post to the mailing list and I'll do
my best to help you debug.

Either way, I think this is a good milestone and I'll be adding the patches
and the regression test for ALPHA_FS/10.linux_boot/InOrder-CPU within the
next week (or two) and then eventually updating more things in our
handy-dandy M5 status matrix:
http://m5sim.org/wiki/index.php/Status_Matrix

The next big thing I want to get in InOrder before the gem5 official merge
is complete is timing TLB translation and in the longer term microcode
support. I can't say exactly when those will be added (again, not enough
time for me to completely focus on this), but if anyone wants to try I have
a good idea of what needs to be done and can assist.

   


___
m5-dev mailing list
m5-dev@m5sim.org
http://m5sim.org/mailman/listinfo/m5-dev


Re: [m5-dev] Introducing...InOrder AlphaFS!

2011-04-21 Thread Steve Reinhardt
Way to go, Korey!

On Thu, Apr 21, 2011 at 1:11 AM, Gabe Black gbl...@eecs.umich.edu wrote:
 Tada! Congratulations.

 Gabe

 On 4/21/2011 12:12 AM, Korey Sewell wrote:

 For 1-4 wide, InOrder Cores running in ALPHA FS mode... Observe terminal
 output:
 M5 console: m5AlphaAccess @ 0xFD02
 Got Configuration 623
 memsize 800 pages 4000
 First free page after ROM 0xFC018000
 HWRPB 0xFC018000 l1pt 0xFC04 l2pt 0xFC042000
 l3pt_rpb 0xFC044000 l3pt_kernel 0xFC048000 l2reserv
 0xFC046000
 kstart = 0xFC31, kend = 0xFC855898, kentry =
 0xFC31, numCPUs = 0x1
 CPU Clock at 2000 MHz IntrClockFrequency=1024
 Booting with 1 processor(s)
 ...
 Linux version 2.6.13 (h...@zed.eecs.umich.edu) (gcc version 3.4.3) #1 SMP
 Sun Oct 8 ..
 Brought up 1 CPUs
 ...
 init started:  BusyBox v1.1.0 (2007.03.04-01:07+) multi-call binary
 mounting filesystems...
 EXT2-fs warning: checktime reached, running e2fsck is recommended
 loading script...
 *
 **Two Words: Yes Sir!

 *More Than Two Words: InOrder is booting Linux!!!

 Getting to this point took about30 patches, so I have a lot of
 patch-cleaning work to do before I can place it in the repo.

 Also, just because it boots doesn't mean I've tested it with any kind of
 real workload suite (e.g. SPEC or Parsec).  I don't have the bandwidth to
 complete stress test through all the benchmarks and test but if someone
 tries and something goes wrong, please post to the mailing list and I'll
 do
 my best to help you debug.

 Either way, I think this is a good milestone and I'll be adding the
 patches
 and the regression test for ALPHA_FS/10.linux_boot/InOrder-CPU within the
 next week (or two) and then eventually updating more things in our
 handy-dandy M5 status matrix:
 http://m5sim.org/wiki/index.php/Status_Matrix

 The next big thing I want to get in InOrder before the gem5 official
 merge
 is complete is timing TLB translation and in the longer term microcode
 support. I can't say exactly when those will be added (again, not enough
 time for me to completely focus on this), but if anyone wants to try I
 have
 a good idea of what needs to be done and can assist.



 ___
 m5-dev mailing list
 m5-dev@m5sim.org
 http://m5sim.org/mailman/listinfo/m5-dev

___
m5-dev mailing list
m5-dev@m5sim.org
http://m5sim.org/mailman/listinfo/m5-dev