- Sort Score
- Result 10 results
- Languages All
Results 1 - 2 of 2 for arithmetic (0.51 sec)
-
lib/fips140/v1.0.0.zip
announced length of x is set based on the actual bit size of the input, // ignoring leading zeroes. func (x *Nat) resetToBytes(b []byte) *Nat { x.reset((len(b) + _S - 1) / _S) if err := x.setBytes(b); err != nil { panic("bigmod: internal error: bad arithmetic") } return x.trim() } // trim reduces the size of x to match its value. func (x *Nat) trim() *Nat { // Trim most significant (trailing in little-endian) zero limbs. // We assume comparison with zero (but not the branch) is constant time. for i :=...
Registered: Tue Sep 09 11:13:09 UTC 2025 - Last Modified: Wed Jan 29 15:10:35 UTC 2025 - 635K bytes - Viewed (0) -
RELEASE.md
9 and CUDA 9.1 sometimes does not properly compute the carry bit when decomposing 64-bit address calculations with large offsets (e.g. `load [x + large_constant]`) into 32-bit arithmetic in SASS. As a result, these versions of `ptxas` miscompile most XLA programs which use more than 4GB of temp memory. This results in garbage results and/or `CUDA_ERROR_ILLEGAL_ADDRESS` failures.
Registered: Tue Sep 09 12:39:10 UTC 2025 - Last Modified: Mon Aug 18 20:54:38 UTC 2025 - 740K bytes - Viewed (1)