Bayan shekara guda na cigaba An saki kunshin ginin GCC 12.1, muhimmin sakin farko na sabon reshe na GCC 12.x.
A karkashin sabon tsarin lambar lambar, an yi amfani da nau'in 12.0 yayin haɓakawa, kuma jim kaɗan kafin fito da GCC 12.1, an riga an yi cokali mai yatsa na reshen GCC 13.0, wanda daga ciki za a samar da babban sigar GCC 13.1 na gaba.
GCC 12.1 Manyan Sabbin Fasali
A cikin wannan sabon sigar da aka gabatar, an yi nuni da cewa se kara goyon bayan CTF debugging format, wanda ke ba da ƙaƙƙarfan ajiya na bayanai game da nau'ikan C, alaƙa tsakanin ayyuka, da alamun lalata. Lokacin da aka saka a cikin abubuwan ELF, tsarin yana ba da damar amfani da allunan alamar EFL don guje wa kwafin bayanai.
Baya ga wannan, an lura da cewa aiki ya ci gaba da fadada goyon baya ga ma'auni na C2X da C ++ 23 na gaba don C da C ++ da kuma cewa karfinsu da sassan gwaji na ma'auni C++20 da C++23 sun inganta a cikin C++ Standard Library.
Don gine-gine x86, ƙarin ƙarin kariya daga raunin na'ura mai sarrafawa wanda ya haifar da kisa na umarni bayan ayyukan tsalle-tsalle mara sharadi. Matsalar ta taso ne daga aiwatar da umarnin da aka riga aka tsara nan da nan tare da bin umarnin tsalle a ƙwaƙwalwar ajiya (SLS, Hasashen Layin Madaidaici). Ana ba da shawarar zaɓin "-mharden-sls" don ba da damar kariya.
An kuma haskaka cewa ya ƙara ma'anar don amfani da masu canjin da ba a fara ba zuwa ga ma'aikacin gwaji a tsaye. Ƙara goyan bayan farko don tantance lambar taro akan abubuwan da aka saka na layi. Ingantattun bin diddigin ƙwaƙwalwar ajiya. Sake rubuta lambar don sarrafa kalamai masu canzawa.
An kara Sabbin kira 30 zuwa libgccjit, ɗakin karatu da aka raba don shigar da janareta na lamba cikin wasu matakai da amfani da shi don JIT bytecode zuwa harhada lambar asali.
A gefe guda kuma, an nuna cewa goyon bayan tsarin CO-RE (Haɗa Sau ɗaya - Gudu Ko'ina) zuwa ƙarshen baya don samar da bytecode BPF, wanda yana ba da damar tattara lambar shirye-shiryen eBPF don kernel Linux sau daya kawai kuma yi amfani da loda na musamman na duniya wanda ke daidaita shirin da aka ɗorawa zuwa nau'in kernel da BTF na yanzu (nau'in BPF). CO-RE yana magance matsalar tare da ɗaukar shirye-shiryen eBPF da aka haɗa waɗanda a baya za a iya amfani da su kawai a cikin sigar kernel waɗanda aka gina su don su, tunda matsayin abubuwan da ke cikin tsarin bayanan ya bambanta daga sigar zuwa sigar.
an kara zuwa RISC-V goyon bayan baya don sababbin haɓaka gine-gine na umarni sets zba, zbb, zbc, da zbs, da kuma kari na ISA don vector da scalar cryptographic ayyuka. An bayar da goyan bayan ƙayyadaddun RISC-V ISA 20191213 ta tsohuwa. Ƙara -mtune= tuta-c906 don ba da damar ingantawa don T-HEAD c906 kernels.
Ara goyon baya ga nau'in __int128_t/integer(irin=16) zuwa ga code generation backend don AMD GPUs dangane da GCN microarchitecture. Yana yiwuwa a yi amfani da har zuwa ƙungiyoyin aiki guda 40 a kowace naúrar ƙididdigewa (CU) kuma har zuwa gabas ɗin koyarwa 16 (wavefront, saitin zaren da aka aiwatar a layi ɗaya ta Injin SIMD) kowace ƙungiya. A baya can, gefen umarni ɗaya kawai a kowace CU an yarda.
Alamun "-march", "-mptx" da "-march-map" an ƙara su zuwa bayan NVPTX., An tsara don tsara lambar ta amfani da tsarin tsarin koyarwa na NVIDIA PTX (Parallel Thread Execution). Tallafin da aka aiwatar don PTX ISA sm_53, sm_70, sm_75 da sm_80. Tsoffin gine-gine shine sm_30.
Baya ga wannan, an yi nuni da cewa gyara matsala inda mai tarawa zai yi binciken da bai cancanta ba Maganar dogara ga mai aiki a lokacin ma'anar samfuri maimakon a lokacin gaggawa. Wannan maganin ya dace da halin da ake ciki don maganganun kira masu dogara.
Ya kamata a lura da cewa a ranar 23 ga Mayu, aikin zai yi bikin cika shekaru 35 da kafa GCC na farko. Daga karshe idan kuna sha'awar ƙarin sani game da shi, zaka iya duba bayanan A cikin mahaɗin mai zuwa.