Re: [Gems-users] cache values 0

[Dinagar Raghunathan <draghuna@xxxxxxx>]

Hi Greg/Team,

 

In fact, Ruby Cycle is 0.

 

Profiler Stats

--------------

Elapsed_time_in_seconds: 514

…….

……

 

Ruby_current_time: 1

Ruby_start_time: 1

Ruby_cycles: 0

 

Can I know how to increase it? I also found the below variables have corresponding values. Also, I have tried by giving cpu-switch-time 1 to a higher number.

 

SIMICS_RUBY_MULTIPLIER                      :                          4

OPAL_RUBY_MULTIPLIER                        :                          1

L1CACHE_TRANSITIONS_PER_RUBY_CYCLE          :                         32

L2CACHE_TRANSITIONS_PER_RUBY_CYCLE          :                         32

DIRECTORY_TRANSITIONS_PER_RUBY_CYCLE        :                         32

 

I have one more quick question. From Quick Start Page I learned that “simics>c” after few steps, the simulation starts and it goes for a long time. I manually intercepted it by “Ctrl+C”. Is it usually takes a long time complete simulation and I have to wait till the Simics Stop?? Or manual interception after few minutes is right thing?

 

simics> opal0.sim-step 10000000000

***  Simics getting shaky, switching to 'safe' mode.

Simics (main thread) received a segmentation fault. Will try to recuperate.

simics> c

Warning: Interrupt key pressed more than once.

         Pressing it again will force Simics back to the command prompt.

         This may corrupt the running simulation.

         Try waiting for Simics to stop first.

USER INTERRUPT: Interrupt key pressed multiple times.

Returning to prompt.

simics>

 

Thanks,

Dinagar

 

From: gems-users-bounces@xxxxxxxxxxx [mailto:gems-users-bounces@xxxxxxxxxxx] On Behalf Of Greg Byrd
Sent: Thursday, January 31, 2008 11:27 AM
To: Gems Users
Subject: Re: [Gems-users] cache values 0

 

Looks to me that you only stepped the simulation by one Ruby cycle.

Dinagar Raghunathan wrote:

Hi team,

 

I ran the ruby and opal as per GEMS wiki Quick Start page and I am finding the stat as below.

 

Total_misses: 0

total_misses: 0 [ 0 0 0 0 ]

user_misses: 0 [ 0 0 0 0 ]

supervisor_misses: 0 [ 0 0 0 0 ]

 

instruction_executed: 4 [ 1 1 1 1 ]

simics_cycles_executed: 8 [ 2 2 2 2 ]

cycles_per_instruction: 0 [ 0 0 0 0 ]

misses_per_thousand_instructions: 0 [ 0 0 0 0 ]

 

All L1 and L2 cache stats are displaying 0 and even the Load, Fetch and Store cycles.

 

L1D_cache cache stats:

  L1D_cache_total_misses: 0

  L1D_cache_total_demand_misses: 0

  L1D_cache_total_prefetches: 0

  L1D_cache_total_sw_prefetches: 0

  L1D_cache_total_hw_prefetches: 0

  L1D_cache_misses_per_transaction: 0

  L1D_cache_misses_per_instruction: 0

  L1D_cache_instructions_per_misses: NaN

 

  L1D_cache_request_size: [binsize: log2 max: 0 count: 0 average: NaN |standard deviation: NaN | 0 ]

 

[The above same result is been shown for L1I and L2.]

 

--- L1Cache ---

 - Event Counts -

Load  0

Ifetch  0

Store  0

L1_to_L2  0

L2_to_L1D  0

L2_to_L1I  0

L2_Replacement  0

 

As mentioned in the Quick start, I have ran the simulation for 1processor. Later after seeing this result, I started playing with processor count and other relative values but still I wasn’t able to increase the load. I know I am asking a basic question.

 

So should I got to execute Microbenchmarks or Tester over the GEMS to have the load over the simulator? If so, can I know how simulate with load?

 

Thanks,

Dinagar

_______________________________________________

Gems-users mailing list

Gems-users@xxxxxxxxxxx

https://lists.cs.wisc.edu/mailman/listinfo/gems-users

Use Google to search the GEMS Users mailing list by adding "site:https://lists.cs.wisc.edu/archive/gems-users/" to your search.