ralc88/gem5

This is an read-only mirror of the gem5 simulator. The upstream repository is stored in https://gem5.googlesource.com, code reviews should be submitted to https://gem5-review.googlesource.com/. The mirrors are synchronized every 15 minutes.

gem5's Issues

SEW-related error

I inform you that there seems to be an error in reading the SEW value.
There seems to be an error in "VectorConfig::get_vtype_sew(uint64_t vtype)" function in "src/cpu/vector_engine/vpu/vector_config/vector_config.cc".
If SEW=32, vsew will output 4, if SEW=64, vsew will output 6. There seems to be a problem with shift, so I edit the value and it works normally. (vt(vtype,2,3) --> vt(vtype,3,3))

No performance improvement with increasing vector lanes

I've been coding for conductiong vector operation using your Vector model.
It's a simple code loading 64 x 64 vector table data and adding (8 x 256 times) 64 elements.
In this code, there is not much performance difference between lanes 1 and 8.
I think when we perform vector operation(vfadd_vv_f32m1), it's not allocated in 8-lane.
Please provide a S/W guide which can utilize vector 8 lanes.

// table size 64 x 64
n = 64;																		// element = 64
vl = vsetvl_e32m1(n);															// VLEN = 64 * 32-bit = 2048-bit
zero = (float*)malloc(sizeof(float)*vl);												// 64 * float malloc
for(i = 0; i < 256; ++i) {								
        vfloat32m1_t vop0 = vle32_v_f32m1(zero);									// vfloat32m1_t init 0
        for (j = 0 ; j < 8; ++j) {														// vector add 8 times
            int index_ij = *get2df<int>(index_iter, i, j, 8);									// table index
            vfloat32m1_t vop0_lo = vle32_v_f32m1(get2df<float>(embed, index_ij, 0, 64));		// vector load form table
            vop0 = vfadd_vv_f32m1(vop0, vop0_lo);									// vector add
        }
        vse32_v_f32m1(get2df<float>(output, i, 0, dimension), vop0);						// store data
}

Error running RiVEC Bencmark Suite on this gem5 rvv simulator

RiVEC Benchmark Suite: https://github.com/RALC88/riscv-vectorized-benchmark-suite. (branch rvv-1.0)

I tried axpy, blackscholes and canneal, the benchmarks run well on spike, but not on this rvv simulator. The serial version works well, but not the vector version.
The output (canneal_vector for example):

warn: CheckedInt already exists in allParams. This may be caused by the Python 2.7 compatibility layer.
warn: Enum already exists in allParams. This may be caused by the Python 2.7 compatibility layer.
warn: ScopedEnum already exists in allParams. This may be caused by the Python 2.7 compatibility layer.
gem5 Simulator System.  http://gem5.org
gem5 is copyrighted software; use the --copyright option for details.

gem5 version [DEVELOP-FOR-V20.1]
gem5 compiled Nov 24 2021 11:05:07
gem5 started Nov 26 2021 21:19:00
gem5 executing on 85d9b18d236f, pid 252615
command line: /home/gem5/build/RISCV/gem5.opt ../test_gem5/simple.py

Global frequency set at 1000000000000 ticks per second
warn: No dot file generated. Please install pydot to generate the dot file and pdf.
warn: DRAM device capacity (8192 Mbytes) does not match the address range assigned (512 Mbytes)
warn: Unknown operating system; assuming Linux.
0: system.remote_gdb: listening for remote gdb on port 7000
Beginning simulation!
info: Entering event queue @ 0.  Starting simulation...
PARSEC Benchmark Suite
Threadcount: 1
100 swaps per temperature step
start temperature: 300
netlist filename: /home/riscv-vectorized-benchmark-suite/_canneal/input/100.nets
number of temperature steps: 8
locs created
locs assigned
netlist created. 100 elements.


Initialization took 0.07918000 secs   
panic: Illegal instruction 0xc22025f3 at pc 0x00000000000108d0: 
Memory Usage: 611208 KBytes
Program aborted at tick 85855631000
--- BEGIN LIBC BACKTRACE ---
/home/gem5/build/RISCV/gem5.opt(+0x18069d0)[0x55faed2c89d0]
/home/gem5/build/RISCV/gem5.opt(+0x181a4ee)[0x55faed2dc4ee]
/lib/x86_64-linux-gnu/libpthread.so.0(+0x153c0)[0x7f478780f3c0]
/lib/x86_64-linux-gnu/libc.so.6(gsignal+0xcb)[0x7f47872e418b]
/lib/x86_64-linux-gnu/libc.so.6(abort+0x12b)[0x7f47872c3859]
/home/gem5/build/RISCV/gem5.opt(+0xbc5935)[0x55faec687935]
/home/gem5/build/RISCV/gem5.opt(+0xf0f2f1)[0x55faec9d12f1]
/home/gem5/build/RISCV/gem5.opt(+0xf1105b)[0x55faec9d305b]
/home/gem5/build/RISCV/gem5.opt(+0x162810d)[0x55faed0ea10d]
/home/gem5/build/RISCV/gem5.opt(+0x1623d67)[0x55faed0e5d67]
/home/gem5/build/RISCV/gem5.opt(+0x162348e)[0x55faed0e548e]
/home/gem5/build/RISCV/gem5.opt(+0x180ea4d)[0x55faed2d0a4d]
/home/gem5/build/RISCV/gem5.opt(+0x182eff8)[0x55faed2f0ff8]
/home/gem5/build/RISCV/gem5.opt(+0x182fded)[0x55faed2f1ded]
/home/gem5/build/RISCV/gem5.opt(+0xc2c200)[0x55faec6ee200]
/home/gem5/build/RISCV/gem5.opt(+0xc13bf1)[0x55faec6d5bf1]
/lib/x86_64-linux-gnu/libpython3.8.so.1.0(+0x2a8738)[0x7f4787ac5738]
/lib/x86_64-linux-gnu/libpython3.8.so.1.0(_PyEval_EvalFrameDefault+0x8dd8)[0x7f478789af48]
/lib/x86_64-linux-gnu/libpython3.8.so.1.0(_PyEval_EvalCodeWithName+0x8fb)[0x7f47879e7e3b]
/lib/x86_64-linux-gnu/libpython3.8.so.1.0(_PyFunction_Vectorcall+0x94)[0x7f4787ac5114]
/lib/x86_64-linux-gnu/libpython3.8.so.1.0(+0x74d6d)[0x7f4787891d6d]
/lib/x86_64-linux-gnu/libpython3.8.so.1.0(_PyEval_EvalFrameDefault+0x7d86)[0x7f4787899ef6]
/lib/x86_64-linux-gnu/libpython3.8.so.1.0(_PyEval_EvalCodeWithName+0x8fb)[0x7f47879e7e3b]
/lib/x86_64-linux-gnu/libpython3.8.so.1.0(PyEval_EvalCodeEx+0x42)[0x7f47879e81c2]
/lib/x86_64-linux-gnu/libpython3.8.so.1.0(PyEval_EvalCode+0x1f)[0x7f47879e85af]
/lib/x86_64-linux-gnu/libpython3.8.so.1.0(+0x1cfbf1)[0x7f47879ecbf1]
/lib/x86_64-linux-gnu/libpython3.8.so.1.0(+0x25f537)[0x7f4787a7c537]
/lib/x86_64-linux-gnu/libpython3.8.so.1.0(+0x74d6d)[0x7f4787891d6d]
/lib/x86_64-linux-gnu/libpython3.8.so.1.0(_PyEval_EvalFrameDefault+0x12fd)[0x7f478789346d]
/lib/x86_64-linux-gnu/libpython3.8.so.1.0(_PyEval_EvalCodeWithName+0x8fb)[0x7f47879e7e3b]
/lib/x86_64-linux-gnu/libpython3.8.so.1.0(_PyFunction_Vectorcall+0x94)[0x7f4787ac5114]
/lib/x86_64-linux-gnu/libpython3.8.so.1.0(+0x74d6d)[0x7f4787891d6d]
--- END LIBC BACKTRACE ---
Aborted (core dumped)

Environment setup:
OS: Ubuntu20.04
llvm: https://github.com/plctlab/llvm-project (branch rvv-iscas)
riscv-gnu-toolchain: https://github.com/riscv-collab/riscv-gnu-toolchain (Newlib)
gem5: https://github.com/RALC88/gem5 (branch develop)

Is it a matter of rvv version? Does this simulator support rvv-1.0?
Or, is the error caused by some other reasons?

warn: No functional unit for OpClass VectorArith1Src & etc.

OS WSL2 Ubuntu-20.04
Branch used develop
Python Python2.7.18

Good day. My friends and I met these warnings while testing the blackscholes benchmark. The warnings said that some of the vector ops aren't available

warn: CheckedInt already exists in allParams. This may be caused by the Python 2.7 compatibility layer.
warn: Enum already exists in allParams. This may be caused by the Python 2.7 compatibility layer.
warn: ScopedEnum already exists in allParams. This may be caused by the Python 2.7 compatibility layer.
gem5 Simulator System.  http://gem5.org
gem5 is copyrighted software; use the --copyright option for details.

gem5 version [DEVELOP-FOR-V20.1]
gem5 compiled Jan  6 2022 17:53:18
gem5 started Jan  7 2022 10:01:00
gem5 executing on DESKTOP-9F2JKA5, pid 1176
command line: /mnt/d/AC/gem5/build/RISCV/gem5.opt /mnt/d/AC/gem5/configs/example/riscv_vector_engine.py '--cmd=/mnt/d/AC/gem5/riscv-vectorized-benchmark-suite/_jacobi-2d/bin/jacobi-2d_vector.exe 1 /mnt/d/AC/gem5/riscv-vectorized-benchmark-suite/_jacobi-2d/input/in_256.input output_vector.txt'

Global frequency set at 1000000000000 ticks per second
warn: DRAM device capacity (32768 Mbytes) does not match the address range assigned (2048 Mbytes)
warn: Unknown operating system; assuming Linux.
warn: No functional unit for OpClass VectorArith1Src
warn: No functional unit for OpClass VectorArith2Src
warn: No functional unit for OpClass VectorArith3Src
warn: No functional unit for OpClass VectorMaskLogical
warn: No functional unit for OpClass VectorReduction
warn: No functional unit for OpClass VectorConvertIntToFP
warn: No functional unit for OpClass VectorConvertFPToInt
warn: No functional unit for OpClass VectorWConvertFPToInt
warn: No functional unit for OpClass VectorWConvertIntToFP
warn: No functional unit for OpClass VectorWConvertFPToFP
warn: No functional unit for OpClass VectorNConvertFPToInt
warn: No functional unit for OpClass VectorNConvertIntToFP
warn: No functional unit for OpClass VectorNConvertFPToFP
warn: No functional unit for OpClass VectorFPCompare
warn: No functional unit for OpClass VectorIntCompare
warn: No functional unit for OpClass VectorSlideUp
warn: No functional unit for OpClass VectorSlideDown
warn: No functional unit for OpClass VectorToScalar
warn: No functional unit for OpClass VectorMemoryLoad
warn: No functional unit for OpClass VectorMemoryStore
warn: No functional unit for OpClass VectorConfig
0: system.remote_gdb: listening for remote gdb on port 7000

We analyzed the stats.txt, and somehow it showed us these OpClass are used:

system.cpu.op_class_0::VectorArith1Src           1600      0.01%     98.47% # Class of committed instruction
system.cpu.op_class_0::VectorArith2Src         227200      0.96%     99.43% # Class of committed instruction
system.cpu.op_class_0::VectorArith3Src          62400      0.26%     99.69% # Class of committed instruction
system.cpu.op_class_0::VectorMaskLogical            0      0.00%     99.69% # Class of committed instruction
system.cpu.op_class_0::VectorReduction              0      0.00%     99.69% # Class of committed instruction
system.cpu.op_class_0::VectorConvertIntToFP         6400      0.03%     99.72% # Class of committed instruction
system.cpu.op_class_0::VectorConvertFPToInt         9600      0.04%     99.76% # Class of committed instruction
system.cpu.op_class_0::VectorWConvertFPToInt            0      0.00%     99.76% # Class of committed instruction
system.cpu.op_class_0::VectorWConvertIntToFP            0      0.00%     99.76% # Class of committed instruction
system.cpu.op_class_0::VectorWConvertFPToFP            0      0.00%     99.76% # Class of committed instruction
system.cpu.op_class_0::VectorNConvertFPToInt            0      0.00%     99.76% # Class of committed instruction
system.cpu.op_class_0::VectorNConvertIntToFP            0      0.00%     99.76% # Class of committed instruction
system.cpu.op_class_0::VectorNConvertFPToFP            0      0.00%     99.76% # Class of committed instruction
system.cpu.op_class_0::VectorFPCompare          11200      0.05%     99.81% # Class of committed instruction
system.cpu.op_class_0::VectorIntCompare          1600      0.01%     99.82% # Class of committed instruction
system.cpu.op_class_0::VectorSlideUp                0      0.00%     99.82% # Class of committed instruction
system.cpu.op_class_0::VectorSlideDown              0      0.00%     99.82% # Class of committed instruction
system.cpu.op_class_0::VectorToScalar               0      0.00%     99.82% # Class of committed instruction
system.cpu.op_class_0::VectorMemoryLoad          9600      0.04%     99.86% # Class of committed instruction
system.cpu.op_class_0::VectorMemoryStore         1600      0.01%     99.86% # Class of committed instruction
system.cpu.op_class_0::VectorConfig             32000      0.14%    100.00% # Class of committed instruction

Questions
Should we ignore the warnings above?
If yes, is the data shown in stats.txt reliable?
Else, what could we do to solve these warnings?