Commentary updates to SHA for sparcv9.

# ====================================================================

# Performance improvement is not really impressive on pre-T1 CPU: +8%

# over Sun C and +25% over gcc [3.3]. While on T1, ... And there

# is a gimmick. X[16] vector is packed to 8 64-bit registers and as

# result nothing is spilled on stack. In addition input data is loaded

# in compact instruction sequence, thus minimizing the window when the

# code is subject to [inter-thread] cache-thrashing hazard. The goal

# is to ensure scalability on UltraSPARC T1, or rather to avoid decay

# when amount of active threads exceeds the number of physical cores.

# over Sun C and +25% over gcc [3.3]. While on T1, a.k.a. Niagara, it

# turned to be 40% faster than 64-bit code generated by Sun C 5.8 and

# >2x than 64-bit code generated by gcc 3.4. And there is a gimmick.

# X[16] vector is packed to 8 64-bit registers and as result nothing

# is spilled on stack. In addition input data is loaded in compact

# instruction sequence, thus minimizing the window when the code is

# subject to [inter-thread] cache-thrashing hazard. The goal is to

# ensure scalability on UltraSPARC T1, or rather to avoid decay when

# amount of active threads exceeds the number of physical cores.

$bits=32;

for (@ARGV)	{ $bits=64 if (/\-m64/ || /\-xarch\=v9/); }

#

# SHA512 on UltraSPARC T1.

#

# ...

# It's not any faster than 64-bit code generated by Sun C 5.8. This is

# because 64-bit code generator has the advantage of using 64-bit

# loads to access X[16], which I consciously traded for 32-/64-bit ABI

# duality [as per above]. But it surpasses 32-bit Sun C generated code

# by 60%, not to mention that it doesn't suffer from severe decay when

# running 4 times physical cores threads and that it leaves gcc [3.4]

# behind by over 4x factor! If compared to SHA256, single thread

# performance is only 10% better, but overall throughput for maximum

# amount of threads for given CPU exceeds corresponding one of SHA256

# by 30% [again, optimal coefficient is 50%].

$bits=32;

for (@ARGV)	{ $bits=64 if (/\-m64/ || /\-xarch\=v9/); }