[Gems-users] Low IPC in Opal

["Daniel Sánchez Pedreño" <sanatox@xxxxxxxxx>]

Dear list,

I have experimented a very low IPC when using Opal in GEMS 2.1. Around 0,6 when the processor width is 4. For example, for the FFT application from SPLASH-2 using just 1 processor, the results are:

Total number of instructions                         57450293
Total number of cycles                               89828116
number of continue calls                             57450293
Instruction per cycle:                             0.639558

>
I have also seen that the fetch stage is usually stalled because of window full event, which represents around the 50% of the number of cycles of the simulation:

Reasons for fetch stalls:
Fetch ready         :                        0   0.00%
Fetch i-cache miss  :                1,553,619   2.73%
Fetch squash        :                      167   0.00%
Fetch I-TLB miss    :                   20,351   0.04%
Window Full         :               43,860,074  77.14% 
Fetch Barrier       :               11,421,297  20.09%
Write Buffer Full   :                        0   0.00%

However, the ROB size in this experiment is 1024 entries while the window size is 512. Additionally,The L1 cache miss ratio is 3%.

Finally, the Retire stage is stalled because of the following events: 

Retire Not-Ready Stage Histogram
FETCH_STAGE   = 201993   ( 0.249%)
DECODE_STAGE   = 202944   ( 0.250%)
READY_STAGE   = 8931472   (11.022%)
EXECUTE_STAGE   = 35066395   (43.273%)
CACHE_MISS_STAGE   = 23505376   (29.006%)
CACHE_NOTREADY_STAGE   = 4287931   ( 5.291%)
COMPLETE_STAGE   = 8839216   (10.908%)

So the question is: is correct this low IPC? From the results obtained, what can I do to increase performance?. 

Thank you.