RFR: JDK-8320448 Accelerate IndexOf using AVX2 [v2]

Sandhya Viswanathan sviswanathan at openjdk.org
Wed Dec 20 22:35:46 UTC 2023


On Wed, 29 Nov 2023 15:01:32 GMT, Scott Gibbons <sgibbons at openjdk.org> wrote:

>> Re-write the IndexOf code without the use of the pcmpestri instruction, only using AVX2 instructions.  This change accelerates String.IndexOf on average 1.3x for AVX2.  The benchmark numbers:
>> 
>> 
>> Benchmark	                                               Score		Latest		
>> StringIndexOf.advancedWithMediumSub   343.573		317.934		0.925375393x
>> StringIndexOf.advancedWithShortSub1	  1039.081		1053.96		1.014319384x
>> StringIndexOf.advancedWithShortSub2	      55.828		110.541		1.980027943x
>> StringIndexOf.constantPattern	                9.361		11.906		1.271872663x
>> StringIndexOf.searchCharLongSuccess	        4.216		4.218		1.000474383x
>> StringIndexOf.searchCharMediumSuccess	3.133		3.216		1.02649218x
>> StringIndexOf.searchCharShortSuccess	3.76		        3.761		1.000265957x
>> StringIndexOf.success	                                9.186		9.713		1.057369911x
>> StringIndexOf.successBig	                      14.341		46.343		3.231504079x
>> StringIndexOfChar.latin1_AVX2_String	  6220.918		12154.52		1.953814533x
>> StringIndexOfChar.latin1_AVX2_char	  5503.556		5540.044		1.006629895x
>> StringIndexOfChar.latin1_SSE4_String	  6978.854		6818.689		0.977049957x
>> StringIndexOfChar.latin1_SSE4_char	  5657.499		5474.624		0.967675646x
>> StringIndexOfChar.latin1_Short_String	  7132.541		6863.359		0.962260014x
>> StringIndexOfChar.latin1_Short_char	16013.389	      16162.437		1.009307711x
>> StringIndexOfChar.latin1_mixed_String	  7386.123	      14771.622		1.999915517x
>> StringIndexOfChar.latin1_mixed_char	  9901.671		9782.245		0.987938803
>
> Scott Gibbons has updated the pull request incrementally with one additional commit since the last revision:
> 
>   Only use optimization when EnableX86ECoreOpts is true

src/hotspot/cpu/x86/stubGenerator_x86_64_string.cpp line 42:

> 40: // 2. Broadcast the last byte of the needle to a different ymm register
> 41: // 3. Compare the first-byte ymm register to the first 32 bytes of the haystack
> 42: // 4. Compare the last-byte register to the 32 bytes of the haystack at the (k-1)st position

It would be good to mention that k is the length of the needle.

src/hotspot/cpu/x86/stubGenerator_x86_64_string.cpp line 117:

> 115: /******************************************************************************/
> 116: 
> 117: void StubGenerator::loop_helper(int size, Label& bailout, Label& loop_top) {

Let us name this method as string_indexof_loop_helper() instead of just loop_helper().

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

PR Review Comment: https://git.openjdk.org/jdk/pull/16753#discussion_r1409626570
PR Review Comment: https://git.openjdk.org/jdk/pull/16753#discussion_r1409623608


More information about the core-libs-dev mailing list