Mercurial > hg > truffle
comparison src/cpu/x86/vm/x86_32.ad @ 1575:3657cb01ffc5
6954029: Improve implicit null check generation with compressed oops
Summary: Hoist DecodeN instruction above null check
Reviewed-by: never, twisti
author | kvn |
---|---|
date | Wed, 02 Jun 2010 09:49:32 -0700 |
parents | 110501f54a99 |
children | e9ff18c4ace7 |
comparison
equal
deleted
inserted
replaced
1574:1eb493f33423 | 1575:3657cb01ffc5 |
---|---|
1374 | 1374 |
1375 // Should the Matcher clone shifts on addressing modes, expecting them to | 1375 // Should the Matcher clone shifts on addressing modes, expecting them to |
1376 // be subsumed into complex addressing expressions or compute them into | 1376 // be subsumed into complex addressing expressions or compute them into |
1377 // registers? True for Intel but false for most RISCs | 1377 // registers? True for Intel but false for most RISCs |
1378 const bool Matcher::clone_shift_expressions = true; | 1378 const bool Matcher::clone_shift_expressions = true; |
1379 | |
1380 bool Matcher::narrow_oop_use_complex_address() { | |
1381 ShouldNotCallThis(); | |
1382 return true; | |
1383 } | |
1384 | |
1379 | 1385 |
1380 // Is it better to copy float constants, or load them directly from memory? | 1386 // Is it better to copy float constants, or load them directly from memory? |
1381 // Intel can load a float constant from a direct address, requiring no | 1387 // Intel can load a float constant from a direct address, requiring no |
1382 // extra registers. Most RISCs will have to materialize an address into a | 1388 // extra registers. Most RISCs will have to materialize an address into a |
1383 // register first, so they would do better to copy the constant from stack. | 1389 // register first, so they would do better to copy the constant from stack. |