- Modify the GCC version used for CI testing of the RISCV architecture

[zijianli1234]

The previous GCC version for RISCV CI was 13, but according to the link, GCC 14 supports RVV 1.0, so change the GCC version for CI testing.
Fix a bug on the ZSTD_row_getRVVMask() function.
Improve some performance by adding RVV optimization to the ZSTD_copy16 function.
After modification:
[root@musecard-00-oerv tests]# ./fullbench enwiki5.txt

	After modification:	before modification:	([After modification]-[before modification]/[before modification]
compress	22.6 MB/s ( 40160)	21.6 MB/s ( 40160)	4%
decompress	123.4 MB/s ( 100811)	115.0 MB/s ( 100811)	7%
compress_freshCCtx	22.1 MB/s ( 40160)	21.9 MB/s ( 40160)	0.9%
decompressDCtx	123.2 MB/s ( 100811)	114.9 MB/s ( 100811)	7%
compressContinue	22.2 MB/s ( 40160)	21.9 MB/s ( 40160)	1%
compressContinue_extDict	18.2 MB/s ( 40150)	18.3 MB/s ( 40150)	-0.5%
decompressContinue	123.1 MB/s ( 100811)	114.3 MB/s ( 100811)	7%
compressStream	20.9 MB/s ( 40157)	19.9 MB/s ( 40157)	5%
compressStream_freshCCtx	20.0 MB/s ( 40157)	20.4 MB/s ( 40157)	-1%
decompressStream	123.1 MB/s ( 100811)	114.8 MB/s ( 100811)	7%
compress2	22.4 MB/s ( 40160)	21.9 MB/s ( 40160)	2%
compressStream2, end	22.5 MB/s ( 40160)	22.3 MB/s ( 40160)	0.8%
compressStream2, end & short	20.5 MB/s ( 40160)	20.4 MB/s ( 40160)	0.4%
compressStream2, continue	20.3 MB/s ( 40157)	20.5 MB/s ( 40157)	-0.4%
compressStream2, -T2, continue	18.3 MB/s ( 40157)	18.2 MB/s ( 40157)	5%
compressStream2, -T2, end	22.5 MB/s ( 40160)	21.8 MB/s ( 40160)	3%
compressSequences	77.2 MB/s ( 40047)	74.4 MB/s ( 40047)	3%
compressSequencesAndLiterals	95.6 MB/s ( 40047)	90.7 MB/s ( 40047)	5%
convertSequences (1st block)	3738.2 MB/s ( 7193)	3746.8 MB/s ( 7193)	-0.2%
get1BlockSummary (1st block)	4732.1 MB/s ( 7193)	4712.7 MB/s ( 7193)	0.4%
decodeLiteralsHeader (1st block:	9415.3 MB/s ( 81)	9264.3 MB/s ( 81)	1.6%
decodeLiteralsBlock (1st block)	367.3 MB/s ( 21537)	371.7 MB/s ( 21537)	-1%
decodeSeqHeaders (1st block)	2083.7 MB/s ( 62)	1861.0 MB/s ( 62)	11%