199 commits

Author	SHA1	Message	Date
Hai Zhu	5b5ee3c4bf	gh-134584: Eliminate redundant refcounting from _LOAD_ATTR_WITH_HINT (GH-143062) ... Eliminate redundant refcounting from _LOAD_ATTR_WITH_HINT	2025-12-23 00:28:08 +00:00
Savannah Ostrowski	1391ee664c	GH-134584: Remove redundant refcount for BINARY_OP_SUBSCR_STR_INT (#142844)	2025-12-18 21:29:54 +00:00
LloydZ	33d94abafd	gh-134584: Eliminate redundant refcounting from _BINARY_OP_SUBSCR_LIST_INT (GH-142926)	2025-12-18 18:25:36 +00:00
Donghee Na	71a7cb8887	gh-134584: Remove redundant refcount from _BINARY_OP_ADD_UNICODE (gh-142825)	2025-12-18 21:33:18 +09:00
Nadeshiko Manju	4345253981	gh-134584: Eliminate redundant refcounting from _STORE_ATTR_WITH_HINT (GH-142767) ... Signed-off-by: Manjusaka <me@manjusaka.me>	2025-12-16 22:21:04 +00:00
Nadeshiko Manju	6ee51a36b3	gh-134584: Eliminate redundant refcounting from _LOAD_ATTR_INSTANCE_VALUE (GH-142769) ... Signed-off-by: Manjusaka <me@manjusaka.me>	2025-12-16 20:39:20 +00:00
Savannah Ostrowski	bef63d2fb8	GH-134584: Remove redundant refcount from _STORE_ATTR_SLOT (#142729)	2025-12-15 15:18:44 -08:00
Ken Jin	790a46a449	gh-134584: JIT: Eliminate redundant refcount ops for X_INT (GH-142765)	2025-12-15 22:22:18 +00:00
Nadeshiko Manju	059316ade6	gh-134584: Eliminate redundant refcounting from _STORE_ATTR_INSTANCE_VALUE (GH-142759) ... Signed-off-by: Manjusaka <me@manjusaka.me>	2025-12-15 20:17:12 +00:00
Ken Jin	0ac4e6c6cd	gh-134584: Remove custom float decref ops (GH-142576)	2025-12-15 19:38:58 +00:00
Donghee Na	abaaeee6a0	gh-134584: Eliminate redundant refcounting from _STORE_SUBSCR_DICT (GH-142712) ... Co-authored-by: Ken Jin <kenjin4096@gmail.com>	2025-12-15 16:42:32 +00:00
Nadeshiko Manju	872ab51f32	gh-134584: Eliminate redundant refcounting from _CALL_LIST_APPEND (GH-142711) ... Signed-off-by: Manjusaka <me@manjusaka.me> Co-authored-by: Ken Jin <kenjin4096@gmail.com>	2025-12-15 15:49:36 +00:00
Donghee Na	04da416e6b	gh-134584: Eliminate redundant refcounting from _STORE_SUBSCR_LIST_INT (gh-142703)	2025-12-14 12:26:00 +00:00
PuQing	a2a400af1e	gh-134584: Eliminate redundant refcounting from _CALL_BUILTION_O (GH-142695) ... Co-authored-by: Ken Jin <kenjin4096@gmail.com>	2025-12-14 10:28:02 +00:00
Nadeshiko Manju	a154c9ed4e	gh-134584: Eliminate redundant refcounting from _CALL_STR_1 (GH-136070) ... Signed-off-by: Manjusaka <me@manjusaka.me>	2025-12-14 09:33:05 +00:00
Ken Jin	e02a35c365	gh-134584: Cleanups for GH-135860 (GH-142604)	2025-12-13 14:38:10 +00:00
Ken Jin	a3a611b042	gh-134584: Revert partially GH-135860 (GH-142620)	2025-12-12 14:04:11 +00:00
Donghee Na	a27538540e	gh-134584: Eliminate redundant refcounting from `_CALL_LEN` (gh-136104)	2025-12-11 15:24:34 +00:00
Noam Cohen	a78f43b001	gh-134584: Eliminate redundant refcounting from _CALL_TUPLE_1 (GH-135860)	2025-12-11 14:31:28 +00:00
Mark Shannon	469f191a85	GH-135379: Top of stack caching for the JIT. (GH-135465) ... Uses three registers to cache values at the top of the evaluation stack This significantly reduces memory traffic for smaller, more common uops.	2025-12-11 10:32:52 +00:00
Mark Shannon	b420f6be53	GH-139109: Support switch/case dispatch with the tracing interpreter. (GH-141703)	2025-11-18 13:31:48 +00:00
Ken Jin	4fa80ce74c	gh-139109: A new tracing JIT compiler frontend for CPython (GH-140310) ... This PR changes the current JIT model from trace projection to trace recording. Benchmarking: better pyperformance (about 1.7% overall) geomean versus current https://raw.githubusercontent.com/facebookexperimental/free-threading-benchmarking/refs/heads/main/results/bm-20251108-3.15.0a1%2B-7e2bc1d-JIT/bm-20251108-vultr-x86_64-Fidget%252dSpinner-tracing_jit-3.15.0a1%2B-7e2bc1d-vs-base.svg, 100% faster Richards on the most improved benchmark versus the current JIT. Slowdown of about 10-15% on the worst benchmark versus the current JIT. **Note: the fastest version isn't the one merged, as it relies on fixing bugs in the specializing interpreter, which is left to another PR**. The speedup in the merged version is about 1.1%. https://raw.githubusercontent.com/facebookexperimental/free-threading-benchmarking/refs/heads/main/results/bm-20251112-3.15.0a1%2B-f8a764a-JIT/bm-20251112-vultr-x86_64-Fidget%252dSpinner-tracing_jit-3.15.0a1%2B-f8a764a-vs-base.svg Stats: 50% more uops executed, 30% more traces entered the last time we ran them. It also suggests our trace lengths for a real trace recording JIT are too short, as a lot of trace too long aborts https://github.com/facebookexperimental/free-threading-benchmarking/blob/main/results/bm-20251023-3.15.0a1%2B-eb73378-CLANG%2CJIT/bm-20251023-vultr-x86_64-Fidget%252dSpinner-tracing_jit-3.15.0a1%2B-eb73378-pystats-vs-base.md . This new JIT frontend is already able to record/execute significantly more instructions than the previous JIT frontend. In this PR, we are now able to record through custom dunders, simple object creation, generators, etc. None of these were done by the old JIT frontend. Some custom dunders uops were discovered to be broken as part of this work gh-140277 The optimizer stack space check is disabled, as it's no longer valid to deal with underflow. Pros: * Ignoring the generated tracer code as it's automatically created, this is only additional 1k lines of code. The maintenance burden is handled by the DSL and code generator. * `optimizer.c` is now significantly simpler, as we don't have to do strange things to recover the bytecode from a trace. * The new JIT frontend is able to handle a lot more control-flow than the old one. * Tracing is very low overhead. We use the tail calling interpreter/computed goto interpreter to switch between tracing mode and non-tracing mode. I call this mechanism dual dispatch, as we have two dispatch tables dispatching to each other. Specialization is still enabled while tracing. * Better handling of polymorphism. We leverage the specializing interpreter for this. Cons: * (For now) requires tail calling interpreter or computed gotos. This means no Windows JIT for now :(. Not to fret, tail calling is coming soon to Windows though https://github.com/python/cpython/pull/139962 Design: * After each instruction, the `record_previous_inst` function/label is executed. This does as the name suggests. * The tracing interpreter lowers bytecode to uops directly so that it can obtain "fresh" values at the point of lowering. * The tracing version behaves nearly identical to the normal interpreter, in fact it even has specialization! This allows it to run without much of a slowdown when tracing. The actual cost of tracing is only a function call and writes to memory. * The tracing interpreter uses the specializing interpreter's deopt to naturally form the side exit chains. This allows it to side exit chain effectively, without repeating much code. We force a re-specializing when tracing a deopt. * The tracing interpreter can even handle goto errors/exceptions, but I chose to disable them for now as it's not tested. * Because we do not share interpreter dispatch, there is should be no significant slowdown to the original specializing interpreter on tailcall and computed got with JIT disabled. With JIT enabled, there might be a slowdown in the form of the JIT trying to trace. * Things that could have dynamic instruction pointer effects are guarded on. The guard deopts to a new instruction --- `_DYNAMIC_EXIT`.	2025-11-13 18:08:32 +00:00
Peter Bierma	b07a267953	gh-137883: Check the recursion limit for specialized keyword argument calls (GH-137887)	2025-08-19 09:53:38 +01:00
Mark Shannon	af15e1d13e	GH-132532: Add new DSL macros to better declare semantics of exits at ends of instructions/uops. (GH-137098)	2025-08-09 15:41:28 +01:00
Mark Shannon	801cf3fcdd	GH-137276: Don't mark uop as escaping if the escaping call is on an exit branch (GH-137277)	2025-08-02 16:49:34 +01:00
Mark Shannon	9731dd2c8d	GH-135379: Specialize int operations for compact ints only (GH-135668)	2025-06-19 11:10:29 +01:00
Ken Jin	7b15873ed0	gh-135474: Specialize arithmetic only on compact ints (GH-135479) ... Specialize arithmetic only on compact ints. This also makes int operations non-escaping in the JIT and in tier 1.	2025-06-14 17:13:32 +08:00
Mark Shannon	f6f4e8a662	GH-132554: "Virtual" iterators (GH-132555) ... * FOR_ITER now pushes either the iterator and NULL or leaves the iterable and pushes tagged zero * NEXT_ITER uses the tagged int as the index into the sequence or, if TOS is NULL, iterates as before.	2025-05-27 15:59:45 +01:00
Mark Shannon	fbe7b87c06	GH-131688: Mark instructions and uops with DECREF_INPUTS as escaping. (GH-133501)	2025-05-22 09:54:35 +01:00
Mark Shannon	6dcb0fdfe0	GH-134282: Always borrow references LOAD_CONST (GH-134284)	2025-05-20 11:24:11 -04:00
Diego Russo	42d03f3933	GH-131798: Split CALL_LIST_APPEND into several uops (GH-134240)	2025-05-19 15:48:55 -04:00
Dino Viehland	3fa30d9e9c	gh-128045: Syncs w/ latest opcode metadata (#134231) ... Fix opcode metadata	2025-05-19 11:08:50 -04:00
Dino Viehland	cc9add695d	gh-128045: Mark unknown opcodes as deopting to themselves (#128044) ... * Mark unknown opcodes as deopting to themselves	2025-05-19 10:15:16 -04:00
Tomas R.	c492ac7252	GH-131798: Split up and optimize CALL_ISINSTANCE (GH-133339)	2025-05-08 14:26:30 -07:00
Irit Katriel	296cd128bf	Revert "gh-133395: add option for extension modules to specialize BINARY_OP/SUBSCR, apply to arrays (#133396)" (#133498)	2025-05-06 13:12:26 +03:00
Diego Russo	9cc77aaf9d	GH-131798: Split CALL_LEN into several uops (GH-133180)	2025-05-05 14:31:48 -07:00
Irit Katriel	082dbf7788	gh-133395: add option for extension modules to specialize BINARY_OP/SUBSCR, apply to arrays (#133396)	2025-05-05 17:46:56 +01:00
Mark Shannon	ac7d5ba96e	GH-133231: Changes to executor management to support proposed sys._jit module (GH-133287) ... * Track the current executor, not the previous one, on the thread-state. * Batch executors for deallocation to avoid having to constantly incref executors; this is an ad-hoc form of deferred reference counting.	2025-05-04 10:05:35 +01:00
Ken Jin	ddac7ac59a	gh-132744: Check recursion limit in CALL_PY_GENERAL (GH-132746)	2025-05-02 17:36:29 +01:00
Irit Katriel	a4be3bc34f	gh-133258: Fix crash in test_index (GH-133262)	2025-05-01 19:15:53 +02:00
Irit Katriel	5529213d4e	gh-100239: specialize BINARY_OP/SUBSCR for list-slice (#132626)	2025-05-01 10:28:52 +00:00
Lysandros Nikolaou	60202609a2	gh-132661: Implement PEP 750 (#132662) ... Co-authored-by: Lysandros Nikolaou <lisandrosnik@gmail.com> Co-authored-by: Bénédikt Tran <10796600+picnixz@users.noreply.github.com> Co-authored-by: Adam Turner <9087854+AA-Turner@users.noreply.github.com> Co-authored-by: Hugo van Kemenade <1324225+hugovk@users.noreply.github.com> Co-authored-by: Wingy <git@wingysam.xyz> Co-authored-by: Koudai Aono <koxudaxi@gmail.com> Co-authored-by: Dave Peck <davepeck@gmail.com> Co-authored-by: Terry Jan Reedy <tjreedy@udel.edu> Co-authored-by: Paul Everitt <pauleveritt@me.com> Co-authored-by: sobolevn <mail@sobolevn.me>	2025-04-30 11:46:41 +02:00
Jelle Zijlstra	922049b613	gh-130907: Treat all module-level annotations as conditional (#131550)	2025-04-28 06:10:28 -07:00
Tomas R.	08e3389e8c	GH-131798: Split up and optimize CALL_TUPLE_1 in the JIT (GH-132851)	2025-04-24 15:55:03 -07:00
Tomas R.	0a387b311e	GH-131798: Split up and optimize CALL_STR_1 in the JIT (GH-132849)	2025-04-24 12:54:46 -07:00
Tomas R.	a6a3dbb7db	GH-131798: JIT: Split CALL_TYPE_1 into several uops (GH-132419)	2025-04-22 09:30:38 -07:00
Sam Gross	da53660f35	gh-131586: Avoid refcount contention in context managers (gh-131851) ... This avoid reference count contention in the free threading build when calling special methods like `__enter__` and `__exit__`.	2025-04-21 15:54:25 -04:00
Brandt Bucher	20926c73b5	GH-131798: Remove JIT guards for dict, frozenset, list, set, and tuple (GH-132289)	2025-04-09 14:32:21 -07:00
Mark Shannon	7099c75550	GH-131498: Cases generator: manage stacks automatically (GH-132074)	2025-04-04 17:59:36 +01:00
Brandt Bucher	1a9d4a1fb3	GH-131798: Allow the JIT to remove more int/float/str guards (GH-131800)	2025-04-01 15:10:15 -07:00