Set the reference nps for sf_16.1 wrt sf_11

[ppigazzini]

Fishtest with Stockfish 11 (#635) had 1.6 Mnps as reference core nps
and 0.7 Mnps as threshold nps to stop a slow worker.
Set the new reference core nps to 691680 nps according to the 56.77% slowdown
of sf_16.1 wrt sf_11 on the ARCH=x86-64-bmi2 of the Ryzen 7 PRO 7840U.
Set the new threshold for slow worker to 208082 nps according to the 70.27%
slowdown of sf_16.1 wrt sf_11 on the ARCH=x86-64-sse41-popcnt of the
Core i7 3770K.

Stockfish compiled with clang++ 18.1.0
Parallel speedup with bench at depth 13

• ARCH=x86-64-bmi2 (AMD Ryzen 7 PRO 7840U)

sf_base =  3451831 +/-   6037 (95%)
sf_test =  1492232 +/-   3012 (95%)
diff    = -1959598 +/-   4685 (95%)
speedup = -56.770% +/- 0.136% (95%)

• ARCH=x86-84-bmi2 (Intel Xeon E5-2680 v3)

sf_base =  1918881 +/-   1002 (95%)
sf_test =   649033 +/-    675 (95%)
diff    = -1269847 +/-   1154 (95%)
speedup = -66.176% +/- 0.060% (95%)

• ARCH=x86-64-sse41-popcnt (Intel Core i7 3770K)

sf_base =  2099079 +/-   9345 (95%)
sf_test =   623973 +/-   2121 (95%)
diff    = -1475105 +/-   7601 (95%)
speedup = -70.274% +/- 0.362% (95%)

Raise worker version to 230 (also server side)

[ppigazzini]

Data as reference to show:

• difference in speedup between bench at 13 depth (average on 100 iterations) and bench at 20 depth (average on 10 iterations)
• the correction factor can be safely computed composing the speedup for consecutive corrections.

Tables

• AMD Ryzen 7 PRO 7840U - bmi2

depth;iter	vs	speedup %	composed	direct
20;10	11 vs 16	-21.041	0.78959
	16 vs 16.1	-39.773	0.60227
	11 vs 16.1	-53.131	0.475546369	0.46869
13;100	11 vs 16	-28.324	0.71676
	16 vs 16.1	-39.239	0.60761
	11 vs 16.1	-56.77	0.435510544	0.4323

• Intel Xeon E5-2680 v3 - bmi2

depth;iter	vs	speedup %	composed	direct
20;10	11 vs 16	-27.657	0.72343
	16 vs 16.1	-49.305	0.50695
	11 vs 16.1	-62.164	0.366742839	0.37836
13;100	11 vs 16	-35.427	0.64573
	16 vs 16.1	-47.327	0.52673
	11 vs 16.1	-66.176	0.340125363	0.33824

• Intel Core i7 3770K - sse41-popcnt

depth;iter	vs	speedup %	composed	direct
20;10	11 vs 16	-34.159	0.65841
	16 vs 16.1	-51.062	0.48938
	11 vs 16.1	-67.457	0.322212686	0.32543
13;100	11 vs 16	-40.87	0.5913
	16 vs 16.1	-50.08	0.4992
	11 vs 16.1	-70.274	0.29517696	0.29726