How big is the level 1 instruction cache?
If you unroll too much, the cache might not keep up with a long run of branchless code, and the branch (when it does come) will be another cache miss as...
Type: Posts; User: Salem
How big is the level 1 instruction cache?
If you unroll too much, the cache might not keep up with a long run of branchless code, and the branch (when it does come) will be another cache miss as...
Re Nominal Animal's post #10
dsub r5, r5, r10 ; r5 = r5 - r10
dsll r5, r5, 3 ; r5 = 8*r5: -8, 0, 8, 16, 24, 32, 40.
ld r6, Counter(r5)...
How expensive is division on your machine?
For example, this should have no branches at all (save for the one in the for loop), but may be expensive on the /
counter[ A[i]/1000 ] ++;
Is the...