July 2013 - ghc-tickets

[GHC] #8095: TypeFamilies painfully slow
by GHC 08 Mar '19

08 Mar '19

#8095: TypeFamilies painfully slow -------------------------------------+------------------------------------- Reporter: MikeIzbicki | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler (Type | Version: 7.6.3 checker) | Operating System: Unknown/Multiple Keywords: | Type of failure: Compile-time Architecture: Unknown/Multiple | performance bug Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: 5321 | -------------------------------------+------------------------------------- I'm using the TypeFamilies extension to generate types that are quite large. GHC can handle these large types fine when they are created manually, but when type families get involved, GHC's performance dies. Unlike in ticket #5321, using tail recursion does not eliminate the problem, and the order of arguments greatly affects compile time. I've attached a file Types.hs that demonstrates the problems. This file generates another Haskell file which has the problems. It takes 3 flags. The first is the size of the type to generate, the second is which type family function to use, and the third is whether to call the type family or just use a manually generated type. Here are my performance results: Using non-tail recursion, I get these results. I have to increase the stack size based on the size of the type I want to generate. {{{ $ ./Types 200 a a > test.hs && time ghc test.hs > /dev/null -fcontext- stack=250 real 0m2.973s $ ./Types 300 a a > test.hs && time ghc test.hs > /dev/null -fcontext- stack=350 real 0m6.018s $ ./Types 400 a a > test.hs && time ghc test.hs > /dev/null -fcontext- stack=450 real 0m9.995s $ ./Types 500 a a > test.hs && time ghc test.hs > /dev/null -fcontext- stack=550 real 0m15.645s }}} Tail recursion generates much slower compile times for some reason, and I still need to adjust the stack size: {{{ $ ./Types 200 b a > test.hs && time ghc test.hs > /dev/null -fcontext- stack=250 real 0m16.120s }}} Changing the order of arguments to the recursive type family greatly changes the run times: {{{ $ ./Types 200 c a > test.hs && time ghc test.hs > /dev/null -fcontext- stack=250 real 0m6.095s }}} Without the type family, I get MUCH better performance: {{{ $ ./Types 10000 a d > test.hs && time ghc test.hs > /dev/null real 0m2.271s }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8095> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 64

[GHC] #8043: Feature Request : Qualified module exports
by GHC 01 Dec '18

01 Dec '18

#8043: Feature Request : Qualified module exports ------------------------------------+------------------------------------- Reporter: erikd | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.7 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: Other Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- I quite like the practice of using `-XNoImplicitPrelude` and then defining my own (often per-project) Prelude doing things like: {{{ module MyProject.Prelude ( module P ) import Data.Text as P (Text) import Prelude as P hiding (head, init, tail, last) }}} As an extension to the above idea, I would like to add qualified module exports defined so that I can do this: {{{ module MyProject.Prelude ( module P , qualified module T ) import Data.Text as P (Text) import Prelude as P hiding (head, init, tail, last) import qualified Data.Text as T }}} so that any file which imports `MyProject.Prelude` will then have all the identifiers that they would normally get if they did: {{{ import qualified Data.Text as T }}} Is this something that would have to go through the Haskell Next procedure? If so, how does one get that kicked off? -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8043> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 13

[GHC] #8089: Implementation of GHC.Event.Poll.poll is broken due to bad coercion
by GHC 16 May '18

16 May '18

#8089: Implementation of GHC.Event.Poll.poll is broken due to bad coercion -------------------------------------------+------------------------------- Reporter: merijn | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: libraries/base | Version: 7.6.3 Keywords: patch | Operating System: Architecture: Unknown/Multiple | Unknown/Multiple Difficulty: Easy (less than 1 hour) | Type of failure: Runtime Blocked By: | crash Related Tickets: | Test Case: | Blocking: -------------------------------------------+------------------------------- Due to an unsafe coercion in GHC.Event.Poll.poll the implementation has a bug that on some platforms (.e.g. linux) results in subtly wrong semantics and on other platforms (e.g. OS X) downright crashes the runtime. The offending line is line 95 in GHC/Event/Poll.hsc: {{{ c_poll ptr (fromIntegral len) (fromIntegral (fromTimeout tout)) }}} `fromTimeout` returns a timeout in seconds as a value of type Int. `c_poll` takes a timeout in seconds as a value of type CInt. If the max value of Int is larger than that of CInt, this coercion results in an overflow, producing a negative value. The faulty semantics are in turn caused by passing this negative value as timeout to `c_poll`. This overflow happens on many 64bit platforms where GHC's Int is 64bit, but CInt is often 32bit. The result of passing a negative value to poll() depends on the platform, but is (on all platforms I checked) different from the desired semantics! On linux any negative timeouts are treated as infinite, this means that a poll() with a timeout that should be finite results in an infinitely blocking poll on linux! On OSX any negative timeout other than -1 (which is treated as infinity) is considered an invalid parameter, returning an error and crashing the runtime. A minimal example reproducing the error (on systems where `fromIntegral (maxBound :: Int) > (fromIntegral (maxBound :: CInt) :: Integer)`): {{{ import Control.Concurrent main = threadDelay maxBound }}} Using 64bit GHC on OSX this produces an immediate runtime crash. The attached patch solves the issue by comparing the Int timeout against the max value of CInt and, if necessary, looping multiple `c_poll` calls (with a timeout equal to `maxBound :: CInt`) if they return `0` (which indicates the timeout has expired with no events occurring). See also: https://github.com/merijn/packages- base/commit/6d8ea02a3f6a2cdb82a9ad786a8c4db780b92091 -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8089> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 22

[GHC] #8091: retry# lacks strictness information
by GHC 13 Sep '17

13 Sep '17

#8091: retry# lacks strictness information ------------------------------------+------------------------------------- Reporter: parcs | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.6.3 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- The `retry#` primop should be marked as returning bottom since it never actually returns anything as far as the simplifier is concerned. This change would help the simplifier eliminate unreachable code inside STM transactions. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8091> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 13

[GHC] #8048: Register spilling produces ineffecient/highly contending code
by GHC 21 May '17

21 May '17

#8048: Register spilling produces ineffecient/highly contending code -------------------------------------+------------------------------------- Reporter: schyler | Owner: Type: bug | Status: new Priority: normal | Milestone: _|_ Component: Compiler | Version: 7.6.3 Keywords: register | Operating System: Unknown/Multiple allocator spill | Type of failure: Runtime Architecture: Unknown/Multiple | performance bug Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | -------------------------------------+------------------------------------- The native codegen and llvm both produce ineffecient code for functions using structures with many strict fields or unboxed values. Consider the following program: {{{ {-# LANGUAGE BangPatterns #-} module Spill where import GHC.Exts data S = S !Int !Int !Int !Int !Int !Int !Int !Int !Int spill :: S -> S -> S -> S spill (S !a !b !c !d !e !f !g !h !i) (S !j !k !l !m !n !o !p !q !r) (S !s !t !u !v !w !x !y !z _) = S (a + j + s) (b + c) (k + r) (a + b + c + d + e + f + g + h + i) (j + k + l + m + n + o + p + q + r) (s + t + u + v + w + x + y + z) (a + b + c) (j + k + l) (s + t + u) }}} Parts of the code produced for this (which is identical regardless of -funbox-strict-fields) looks like: {{{ _cnc: addq $80,%r12 cmpq 144(%r13),%r12 ja _cni movq $Spill.S_con_info,-72(%r12) movq 32(%rbp),%rax movq %rax,-64(%r12) movq 24(%rbp),%rax movq %rax,-56(%r12) movq 16(%rbp),%rax movq %rax,-48(%r12) movq 8(%rbp),%rax movq %rax,-40(%r12) movq 40(%rbp),%rax movq %rax,-32(%r12) movq 48(%rbp),%rax movq %rax,-24(%r12) movq 56(%rbp),%rax movq %rax,-16(%r12) movq 64(%rbp),%rax movq %rax,-8(%r12) movq 7(%rbx),%rax movq %rax,0(%r12) leaq -71(%r12),%rbx addq $72,%rbp jmp *0(%rbp) }}} {{{ _csv: movq 63(%rbx),%rax movq %rax,-56(%rbp) movq 55(%rbx),%rax movq %rax,-48(%rbp) movq 47(%rbx),%rax movq %rax,-40(%rbp) movq 39(%rbx),%rax movq %rax,-32(%rbp) movq 31(%rbx),%rax movq %rax,-24(%rbp) movq 23(%rbx),%rax movq %rax,-16(%rbp) movq 71(%rbx),%rax movq %rax,-8(%rbp) movq 15(%rbx),%rax movq %rax,0(%rbp) }}} And likewise for LLVM: {{{ .LBB10_1: # %coZ movq 7(%rbx), %rcx movq $Spill_S_con_info, 8(%rax) movq 8(%rbp), %rdx movq %rdx, 16(%rax) movq 16(%rbp), %rdx movq %rdx, 24(%rax) movq 24(%rbp), %rdx movq %rdx, 32(%rax) movq 32(%rbp), %rdx movq %rdx, 40(%rax) movq 40(%rbp), %rdx movq %rdx, 48(%rax) movq 48(%rbp), %rdx movq %rdx, 56(%rax) movq 56(%rbp), %rdx movq %rdx, 64(%rax) movq 64(%rbp), %rdx movq %rdx, 72(%rax) movq %rcx, (%r12) movq 72(%rbp), %rax leaq 72(%rbp), %rbp leaq -71(%r12), %rbx jmpq *%rax # TAILCALL }}} Quoting from #ghc "the [register allocator] core algo is '96 vintage". Improvements are needed; * Take into consideration pipelining and handle spills less dramatically, attempting to reduce register contention * Sink memory reads in order to reduce register pressure -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8048> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 16

[GHC] #8082: Ordering of assembly blocks affects performance
by GHC 10 May '17

10 May '17

#8082: Ordering of assembly blocks affects performance ----------------------------------+--------------------------------- Reporter: jstolarek | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler (NCG) | Version: 7.6.3 Keywords: | Operating System: Linux Architecture: x86_64 (amd64) | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ----------------------------------+--------------------------------- During my work on #6135 I noticed that performance of reverse-complem benchmark in nofib depends highly on the order in which assembly block are laid out. With my patches I am consistently getting a 18-20% speed-up. In theory my patches should not impact performance of existing programs, but for some reason they affect the ordering of generated assembly blocks. On of the earlier versions of my patch I noticed that kahan benchmark suffered a 16% performance hit and again the only difference I noticed in the generated assembly was ordering of blocks. I did a more in-depth investigation in case of kahan and it turned out that this difference results from the way Core is generated: the difference between HEAD and my patches was that a worker function had its three parameters passed in different order. I did not investigate this for reverse-complem because Core is considerably larger, but I could spend some time on it if it might be relevant. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8082> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 8

[GHC] #8101: No pattern match non-exhaustiveness warnings when compiling with -fno-code
by GHC 07 May '17

07 May '17

#8101: No pattern match non-exhaustiveness warnings when compiling with -fno-code -------------------------+------------------------------------------------- Reporter: exbb2 | Owner: Type: bug | Status: new Priority: | Milestone: normal | Version: 7.6.3 Component: | Operating System: Unknown/Multiple Compiler | Type of failure: Incorrect warning at Keywords: | compile-time Architecture: | Test Case: Unknown/Multiple | Blocking: Difficulty: | Unknown | Blocked By: | Related Tickets: | -------------------------+------------------------------------------------- {{{ {-# OPTIONS_GHC -Wall #-} module A where data ABC = A | B | C abc :: ABC -> Int abc x = case x of A -> 1 }}} {{{ $ ghc -fno-code -fforce-recomp /tmp/A.hs [1 of 1] Compiling A ( /tmp/A.hs, nothing ) }}} {{{ $ ghc -fobject-code -fforce-recomp /tmp/A.hs [1 of 1] Compiling A ( /tmp/A.hs, /tmp/A.o ) /tmp/A.hs:7:9: Warning: Pattern match(es) are non-exhaustive In a case alternative: Patterns not matched: B C }}} -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8101> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 9

[GHC] #8099: Alternate syntax for indicating when a function is "fully applied" for purposes of inlining
by GHC 04 Aug '16

04 Aug '16

#8099: Alternate syntax for indicating when a function is "fully applied" for purposes of inlining ------------------------------------+------------------------------------- Reporter: jberryman | Owner: Type: feature request | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.6.3 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- I just remembered an old idea I put to the GHC user's list [http://www .mail-archive.com/glasgow-haskell-users(a)haskell.org/msg22119.html here], and thought I should present it here. As I understand it GHC inlines inlines only after a function is fully applied with respect to the syntactic left-hand side, so in order to control inlining and sharing, we sometimes see strange things like: {{{ {-# INLINE foo #-} foo :: a -> b -> c foo a = \b -> ... }}} It would be nice to be able to define the function syntactically in the way that's the most clear (e.g. in pointfree style, etc.) and specify at what point to inline something like this: {{{ foo :: a -> {-# INLINE #-} b -> c foo a b = ... -- or: foo = ... }}} ...which tells the compiler to inline after applying the arguments to the left of the pragma, and GHC would do whatever trivial (I assume?) eta- conversion was required. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8099> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 18

[GHC] #8064: Warning out when "undefined" value has been used.
by GHC 23 Jan '16

23 Jan '16

#8064: Warning out when "undefined" value has been used. ------------------------------------+------------------------------------- Reporter: freizl | Owner: Type: feature request | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.6.3 Keywords: | Operating System: Unknown/Multiple Architecture: Unknown/Multiple | Type of failure: None/Unknown Difficulty: Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | ------------------------------------+------------------------------------- Programmer may like to use "undefined" to define a function or value, which usually means "later implementation". However, people might forget to implement it thus cause run time error. Wish GHC will show some warning/errors at compilation. This idea comes from this article (http://www.devalot.com/articles/2013/07 /why-haskell.html) -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8064> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 1

[GHC] #8060: Undefined symbols when using Template Haskell linked from another object with unexposed modules
by GHC 08 Jan '16

08 Jan '16

#8060: Undefined symbols when using Template Haskell linked from another object with unexposed modules -----------------------------------+--------------------------------------- Reporter: tvynr | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.6.3 Keywords: | Operating System: Linux Architecture: | Type of failure: Compile-time crash Unknown/Multiple | Test Case: Difficulty: Unknown | Blocking: Blocked By: | Related Tickets: | -----------------------------------+--------------------------------------- I have constructed two Cabal projects named "first" and "second". The "first" project defines three files: Logging.hs (which provides some basic logging functionality) and FirstExample.hs and FirstDependent.hs (the latter of which imports the former). In the cabal file, only Logging and FirstDependent are exposed modules. FirstDependent uses a Template Haskell function "loggingFunctions" which defines logging functions for that module. Using cabal or cabal-dev to build this package works quite nicely. The second project depends on the first project and contains a single file, Second.hs. This file imports Logging but neither of the other two files. It also uses the loggingFunctions routine to generate local logging functions. The presence of the "$(loggingFunctions)" splice in Second.hs makes the problem visible, but the real problem is in the first project. Upon compilation of the "second" project with cabal or cabal- dev, the following error is generated: {{{ cabal-dev -s ../cabal-dev-sandbox --with-cabal-install=cabal configure Resolving dependencies... Configuring second-0.1... cabal-dev -s ../cabal-dev-sandbox --with-cabal-install=cabal build Building second-0.1... Preprocessing library second-0.1... [1 of 1] Compiling Second ( lib-src/Second.hs, dist/build/Second.o ) Loading package ghc-prim ... linking ... done. Loading package integer-gmp ... linking ... done. Loading package base ... linking ... done. Loading package array-0.4.0.1 ... linking ... done. Loading package deepseq-1.3.0.1 ... linking ... done. Loading package containers-0.5.0.0 ... linking ... done. Loading package filepath-1.3.0.1 ... linking ... done. Loading package old-locale-1.0.0.5 ... linking ... done. Loading package time-1.4.0.1 ... linking ... done. Loading package bytestring-0.10.0.2 ... linking ... done. Loading package unix-2.6.0.1 ... linking ... done. Loading package directory-1.2.0.1 ... linking ... done. Loading package transformers-0.3.0.0 ... linking ... done. Loading package mtl-2.1.2 ... linking ... done. Loading package text-0.11.3.1 ... linking ... done. Loading package parsec-3.1.3 ... linking ... done. Loading package network-2.4.1.2 ... linking ... done. Loading package process-1.1.0.2 ... linking ... done. Loading package hslogger-1.2.1 ... linking ... done. Loading package pretty-1.1.1.0 ... linking ... done. Loading package template-haskell ... linking ... done. Loading package first-0.2 ... linking ... ghc: /home/some- directories/example/first/../cabal-dev- sandbox//lib/first-0.2/ghc-7.6.1.20121207/HSfirst-0.2.o: unknown symbol `firstzm0zi2_FirstExample_zdfShowErrorSpec3_closure' ghc: unable to load package `first-0.2' make[1]: *** [build] Error 1 }}} Upon further investigation, it would seem that the .o file constructed to represent the "first" project has not been completely constructed: {{{ $ objdump -t cabal-dev- sandbox/lib/first-0.2/ghc-7.6.1.20121207/HSfirst-0.2.o | grep '\*UND\*' | grep first 00000000 *UND* 00000000 firstzm0zi2_FirstExample_zdfOrdErrorSpeczuzdczg_info 00000000 *UND* 00000000 firstzm0zi2_FirstExample_zdfShowErrorSpec3_closure 00000000 *UND* 00000000 firstzm0zi2_FirstExample_zdfOrdErrorSpeczuzdczl_info 00000000 *UND* 00000000 firstzm0zi2_FirstExample_zdfEqErrorSpeczuzdczeze_info 00000000 *UND* 00000000 firstzm0zi2_FirstExample_zdfOrdErrorSpeczuzdczlze_info 00000000 *UND* 00000000 firstzm0zi2_FirstExample_zdfOrdErrorSpeczuzdccompare_info 00000000 *UND* 00000000 firstzm0zi2_FirstExample_zdfOrdErrorSpeczuzdczgze_info }}} (While the above dump is for ghc 7.6.1, this has also been tested on ghc 7.6.3 as well as on both i386- and amd64-based OS installs.) If the "$(loggingFunctions)" splice is removed from Second.hs, then the project builds (presumably because the logic used by non-Template Haskell projects for linking differs from that of Template Haskell projects). Fortunately, it seems that exposing every module of the "first" project resolves this issue. But please note that the FirstExample.hs and FirstDependent.hs files are completely unrelated to the Second.hs file; the latter does not depend on the former in any way. It seems strange that I can't use Template Haskell across projects without exposing all of my project modules. -- Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/8060> GHC <http://www.haskell.org/ghc/> The Glasgow Haskell Compiler

1 4