Confirmed users
753
edits
User:Bjacob/ArithmeticTimingDifferences: Difference between revisions
User:Bjacob/ArithmeticTimingDifferences (view source)
Revision as of 18:26, 20 June 2013
, 20 June 2013→Some Intel CPUs have slow multiplications with NaN values, even with SSE2 instructions
No edit summary |
|||
| Line 126: | Line 126: | ||
Conventional "wisdom" is that while NaN's are slow with x87, they aren't anymore with SSE instructions. That turned out not to be true in full generality. | Conventional "wisdom" is that while NaN's are slow with x87, they aren't anymore with SSE instructions. That turned out not to be true in full generality. | ||
In particular, this is not true for a wide range of Intel CPUs based on the Netburst microarchitecture, which was popular until the mid-2000s. At least Intel family 15 models 2, 3 and 4 are affected: | |||
* [http://people.mozilla.org/~bjacob/ta/results/intel-pentium-4-2.53GHz-family-15-model-2-stepping-9-dell-dimension2400-gavin/time-math-x86-32-sse2-DAZ-FTZ.txt this Pentium 4], family 15 model 2 | |||
* [http://people.mozilla.org/~bjacob/ta/results/intel-pentium-4-3.2GHz-family-15-model-3-stepping-3-tycho/time-math-x86-32-sse2-DAZ-FTZ.txt this Pentium 4], family 15 model 3 | |||
* [http://people.mozilla.org/~bjacob/ta/results/intel-celeron-1.8GHz-family-15-model-2-stepping-7-toshiba-satellite/ this Celeron], family 15 model 2 | |||
* [http://people.mozilla.org/~bjacob/ta/results/intel-pentium-d-3GHz-family-15-model-4-stepping-7-dell-dimension9150-gavin/time-math-x86-64-DAZ-FTZ.txt this Pentium D], family 15 model 4. This set of results shows that the problem exists even in 64bit mode. | |||
=== Some Intel CPUs have abnormally slow comparisons with NaN and denormals, even with SSE2, DAZ and FTZ === | === Some Intel CPUs have abnormally slow comparisons with NaN and denormals, even with SSE2, DAZ and FTZ === | ||