Re: Number of ALU in la464.md

2022-09-26 Thread Xi Ruoyao via Gcc
On Tue, 2022-09-27 at 10:40 +0800, Lulu Cheng wrote: > You are right, I just added it before, but I didn't see the performance > improvement, > > so I didn't change this place. Alright, so just left la464.md as-is for now. I'll try Phoronix bench suite on LoongArch once a binary distro is in a

Re: Number of ALU in la464.md

2022-09-26 Thread Lulu Cheng
You are right, I just added it before, but I didn't see the performance improvement, so I didn't change this place. 在 2022/9/25 下午7:28, Xi Ruoyao 写道: Hi, According to the schematic (https://loongson.github.io/LoongArch-Documentation/Loongson-3A5000-usermanual-EN.html#la464-processor-core, Fig

Number of ALU in la464.md

2022-09-25 Thread Xi Ruoyao via Gcc
Hi, According to the schematic (https://loongson.github.io/LoongArch-Documentation/Loongson-3A5000-usermanual-EN.html#la464-processor-core, Figure 5) we have 4 ALUs in LA464, but in la464.md we only define la464_alu1 and la464_alu2. Should we add la464_alu{3,4} as well, or am I missing something