inefficient Des3DkCrypto/DigestMD5Base.setParityBit()
Hi! While looking at the bitCount() thing I mentioned in the other thread, I noticed that Des3DkCrypto and DigestMD5Base are using a very inefficient implementation of setParityBit(). The one from DESKeyGenerator is much better and uses Integer.bitCount(), which could benefit from a population-count intrinsic. Should I file a CR or even propose a patch? -- Christian
Christian Thalinger wrote:
Hi!
While looking at the bitCount() thing I mentioned in the other thread, I noticed that Des3DkCrypto and DigestMD5Base are using a very inefficient implementation of setParityBit(). The one from DESKeyGenerator is much better and uses Integer.bitCount(), which could benefit from a population-count intrinsic.
Should I file a CR or even propose a patch?
-- Christian
I don't know if the security folks on this mailing list so it might be better to send it to security-dev. -Alan.
On Tue, 2009-02-17 at 15:18 +0000, Alan Bateman wrote:
Christian Thalinger wrote:
Hi!
While looking at the bitCount() thing I mentioned in the other thread, I noticed that Des3DkCrypto and DigestMD5Base are using a very inefficient implementation of setParityBit(). The one from DESKeyGenerator is much better and uses Integer.bitCount(), which could benefit from a population-count intrinsic.
Should I file a CR or even propose a patch?
-- Christian
I don't know if the security folks on this mailing list so it might be better to send it to security-dev.
Ohh, good point, thanks. I thought everything related to the core library should be posted here. I will forward it to the security-dev list. -- Christian
participants (2)
-
Alan Bateman
-
Christian Thalinger