diff options
Diffstat (limited to 'problems')
| -rw-r--r-- | problems/09-Transpose/01.vsm | 8 | ||||
| -rw-r--r-- | problems/09-Transpose/02.vsm | 56 |
2 files changed, 56 insertions, 8 deletions
diff --git a/problems/09-Transpose/01.vsm b/problems/09-Transpose/01.vsm index 223cf2a..3c2e27d 100644 --- a/problems/09-Transpose/01.vsm +++ b/problems/09-Transpose/01.vsm @@ -1,29 +1,21 @@ # (64 / 8).times do |i| # puts "lpassa $lm#{(2*i) * 8}v $ln#{2*i}v16" # puts "lpassa $lm#{(2*i+1) * 8}v $ln#{2*i + 64}v16" -# puts # end lpassa $lm0v $ln0v16 lpassa $lm8v $ln64v16 - lpassa $lm16v $ln2v16 lpassa $lm24v $ln66v16 - lpassa $lm32v $ln4v16 lpassa $lm40v $ln68v16 - lpassa $lm48v $ln6v16 lpassa $lm56v $ln70v16 - lpassa $lm64v $ln8v16 lpassa $lm72v $ln72v16 - lpassa $lm80v $ln10v16 lpassa $lm88v $ln74v16 - lpassa $lm96v $ln12v16 lpassa $lm104v $ln76v16 - lpassa $lm112v $ln14v16 lpassa $lm120v $ln78v16 diff --git a/problems/09-Transpose/02.vsm b/problems/09-Transpose/02.vsm new file mode 100644 index 0000000..cafa0b8 --- /dev/null +++ b/problems/09-Transpose/02.vsm @@ -0,0 +1,56 @@ +imm i"3" $s3/1000 +imm i"7" $s5/1000 +imm i"4" $s7/1000 +imm i"2" $s9/1000 +llsr $peid $ls2 $lr0/1000 # i +land $peid $ls4 $lr2/1000 # j +llsr $aluf $ls8 $lr4/1000 # j1 +land $lr2 $ls2 $lr6/1000 # j2 +lpackbit $aluf $ln0 $nowrite +ladd $aluf $lr4 $nowrite +llsl $aluf $ls2 $nowrite +ladd $aluf $lr0 $nowrite +lpackbit $aluf $ln0 $t +nop +l1bmd $lmt0 $lb0 +nop +nop +l1bmp $llb0 $lln0v +l1bmp $llb8 $lln16v +l1bmp $llb16 $lln32v +l1bmp $llb24 $lln48v +l1bmp $llb32 $lln64v +l1bmp $llb40 $lln80v +l1bmp $llb48 $lln96v +l1bmp $llb56 $lln112v + +16.times do |mabid| + 4.times do |subpeid| + peid = mabid*4 + subpeid + i = peid >> 3 # $lr[0] + j = peid & 0b111 # $lr[2] + j1 = j >> 2 # $lr[4] + j2 = j & 0b11 # $lr[6] + lm = (((j1 + (j2 << 1)) << 3) + i) << 1 + puts "MAB #{mabid} PE #{subpeid} (#{peid}): $lm[#{lm}]" + end +end + + +dvpassa ; t ; x1 +dvpassa ; nop ; x1 +dvpassa ; l1bmd ; x1 +dvpassa ; nop ; x1 + ; nop ; x2 !!! +l1bmp ... ... ; x2 +l1bmp ... ... ; x2 +l1bmp ... ... ; x2 +l1bmp ... ... ; x2 +l1bmp ... ... ; x2 + + + +lpassa $llm0v $llr0v ; dvpassa $llm0v $ln[TODO] +lpassa $llm16v $llr16v ; dvpassa $llm16v $ln[TODO] +dvpassa $lm[TODO] $ln[TODO] +dvpassa $lm[TODO] $ln[TODO] |