A low Overhead Per Object Write Barrier for Smalltalk
- 1. A low Overhead Per Object Write Barrier for the CogVM Clément Béra
- 2. Introduction • The CogVM is the standardVM for: • Pharo • Squeak • Newspeak • Cuis
- 3. Introduction • Working runtime optimizer for Cog’s JIT • Problem with literal mutability
- 4. Problem • Is it possible to mark any object as read- only ? • Smalltalk code to handle mutation failure • Overhead
- 5. Terminology • Discussion onVM mailing-list • Immutable: state cannot change after object’s initialization • Write barrier or read-only object
- 6. Use-cases • Modification tracker • Read-only literals • Compiler optimizations • Inconsistent literal modifications • Others...
- 7. This paper • NOT about framework built using read- only objects • Implementation details to limit the overhead
- 8. Feature • Any object can be marked as read-only, except: • Immediate objects • Context instances • Objects related to Process scheduling • Objects internal to the runtime
- 9. APIs • Object >> isReadOnlyObject • Object >> setIsReadOnlyObject: • Object >> beWritableObject • Object >> beReadOnlyObject
- 10. Read-only object • Instance variable store fail • Primitives mutating a read-only object fail
- 11. IV store failure • Instance variable is not set. • A call-back is sent:
- 12. Primitive failure • First value of array is not set
- 13. Primitive error code • new error code: #'no modification'
- 14. Other details • Support flags • Mirror primitives • Object >> object:setIsReadOnlyObject:
- 15. VM compilation option • VM C compiler flag • TheVM can be compiled with or without the feature.
- 16. Implementation • Object representation • Interpreter support • JIT support
- 17. Implementation • Most critical part: • How to keep IV store efficient ? • Machine code generated by the JIT • Discussed in the paper...
- 18. IV Store details • Wanted • to show it, x86 Assembly Meaning movl -12(%ebp), %edx popl %edi movl %edi, %ds:0x8(%edx) testl 0x00000003, %edi If the value to store is immediate, jump after the store check.jnz after_store Jump after the store check if the receiver is young: compare the young object space limit with receiver address movl 0x00040088, %eax jb after_store cmpl %eax, %edx Load the receiver in %edx. jnb after_store cmpl %eax, %edi If the value to store is an old object, jump after the store check. jnz after_store testb 0x20, %al movzbl %ds:0x3(%edx), %eax If the receiver is already in the remembered table, jump after the store check. call store_check_trampoline Calls the store check trampoline. Perform the store in the first instance variable using both registers (%edx and %edi) Load the value to store in %edi. after_store: Code following the store. x86 Assembly Meaning movl -12(%ebp), %edx popl %ecx movl %ecx, %ds:0x8(%edx) testb 0x03, %cl If the value to store is immediate, jump after the store check.jnz after_store If the receiver is a young object, jump after the store check. movl 0x00040088, %eax jb after_store cmpl %eax, %edx jnb after_store cmpl %eax, %ecx If the value to store is an old object, jump after the store check. jnz after_store testb 0x20, %al movzbl %ds:0x3(%edx), %eax If the receiver is already in the remembered table, jump after the store check. call store_trampoline Calls the store check trampoline. Perform the store in the first instance variable using both registers (%edx and %ecx) Load the receiver in %edx. Load the value to store in %ecx. movl -12(%ebp), %edx movl %ds:(%edx), %eax testl 0x00800000, %eax jnz store_trampoline If the receiver is read-only, jump to the store trampoline. Restore the receiver (to keep its register live). store_trampoline: after_store: Code following the store.
- 19. Evaluation: Slow-down • Binary trees • IV Store intensive • No significant difference
- 20. Evaluation: Slow-down • Pathological case: setter
- 21. Evaluation: Slow-down • At writing time, setter overhead was 17% • Stack frame creation problem • Two path compilation • Now faster than before
- 22. Conclusion • New feature: read-only object • Overhead is very limited