wren/test/benchmark/map_numeric.py

from __future__ import print_function

import time

start = time.clock()

map = {}

for i in range(1, 2000001):
  map[i] = i

sum = 0
for i in range(1, 2000001):
  sum = sum + map[i]
print(sum)

for i in range(1, 2000001):
  del map[i]

print("elapsed: " + str(time.clock() - start))
Adding map benchmark scripts. 2015-02-09 11:51:09 +01:00			`from __future__ import print_function`

			`import time`

			`start = time.clock()`

Add benchmark for string map keys. 2015-02-11 22:41:59 -08:00			`map = {}`

Fix horrendously bad bit hashing function. hashBits() is used to generate a hash code from the same 64 bits used to represent a Wren number as a double. When building a map containing a large number of integer keys, it's important for this to do a good job scattering the bits across the 32-bit key space. Alas, it does not. Worse, the benchmark to test this happens to stop just before the performance falls off a cliff, so this was easy to overlook. This replaces it with the hash function V8 uses, which has much better performance across the numeric range. 2019-07-27 13:34:07 -07:00			`for i in range(1, 2000001):`
Adding map benchmark scripts. 2015-02-09 11:51:09 +01:00			`map[i] = i`

			`sum = 0`
Fix horrendously bad bit hashing function. hashBits() is used to generate a hash code from the same 64 bits used to represent a Wren number as a double. When building a map containing a large number of integer keys, it's important for this to do a good job scattering the bits across the 32-bit key space. Alas, it does not. Worse, the benchmark to test this happens to stop just before the performance falls off a cliff, so this was easy to overlook. This replaces it with the hash function V8 uses, which has much better performance across the numeric range. 2019-07-27 13:34:07 -07:00			`for i in range(1, 2000001):`
Adding map benchmark scripts. 2015-02-09 11:51:09 +01:00			`sum = sum + map[i]`
			`print(sum)`

Fix horrendously bad bit hashing function. hashBits() is used to generate a hash code from the same 64 bits used to represent a Wren number as a double. When building a map containing a large number of integer keys, it's important for this to do a good job scattering the bits across the 32-bit key space. Alas, it does not. Worse, the benchmark to test this happens to stop just before the performance falls off a cliff, so this was easy to overlook. This replaces it with the hash function V8 uses, which has much better performance across the numeric range. 2019-07-27 13:34:07 -07:00			`for i in range(1, 2000001):`
Adding map benchmark scripts. 2015-02-09 11:51:09 +01:00			`del map[i]`

			`print("elapsed: " + str(time.clock() - start))`