Pi computation speed

[TuomLarsen]

Hi!

I read in (1) that FLINT 3 can compute 100 million digits of Pi in 31.6s (8-core AMD Ryzen 7 PRO 5850U CPU (Zen 3)). In the same comparison was y-cruncher which was done in 18.3s.

However I was not able to reproduce such timings so I was hoping if someone could maybe explain what I'm doing wrong.
I'm on Zen 5 CPU and Linux (I disabled multi threading for both FLINT and y-cruncher) and the timings I get are: ~30s for FLINT 3 (compiled from source, uses fft_small, GMP as provided by default in Fedora 43) and ~1s for y-cruncher (-noSMT).

In summary the blog says the difference between FLINT and y-cruncher is ~1.7x but I see ~30x difference.

Thanks for any comments!

(1) https://fredrikj.net/blog/2023/04/flint-furnished-with-faster-fft/

[fredrik-johansson]

Probably y-cruncher is still multithreaded?

According to https://www.numberworld.org/y-cruncher/benchmarks/charts/100m.html the fastest reported y-cruncher timings for 100M are in the 0.5 - 1 second range on various multicore CPUs. These are surely results with multithreading enabled.

[TuomLarsen]

Thank you very much for the reply but no, multithreading was disabled (and didn't make a lot of difference at this size if it was enabled).

The page https://www.numberworld.org/y-cruncher/#Benchmarks also show that the slowest CPU Core i3 8121U computed 100M in ~9s and I would assume the above AMD Ryzen was faster.

[TuomLarsen]

Do you happen to remember which version of y-cruncher was used there?

[fredrik-johansson]

Here's the output benchmarking 100M digits on my machine with the latest version of y-cruncher.

It takes 18.8 seconds single-threaded (first run) and 4.1 seconds multi-threaded (second run).

What kind of output do you get with the same commands?

$ ./y-cruncher 
y-cruncher v0.8.7 Build 9547-gcc

Detecting Environment...

Hardware Features:
(*) Indicates it is used explicitly by y-cruncher.

CPU Vendor:
    AMD         = Yes
    Intel       = No

OS Features:
  * 64-bit      = Yes
  * OS AVX      = Yes
  * OS AVX512   = No

Hardware Features:
    MMX         = Yes
  * x64         = Yes
  * ABM         = Yes
    RDRAND      = Yes
    RDSEED      = Yes
  * BMI1        = Yes
  * BMI2        = Yes
  * ADX         = Yes
    PREFETCHW   = Yes
    PREFETCHWT1 = No
    RDPID       = Yes
    SERIALIZE   = No
    CLDEMOTE    = No

SIMD: 128-bit
  * SSE         = Yes
  * SSE2        = Yes
  * SSE3        = Yes
  * SSSE3       = Yes
    SSE4a       = Yes
  * SSE4.1      = Yes
  * SSE4.2      = Yes
    PCLMULQDQ   = Yes
    AES         = Yes
    SHA         = Yes
    GFNI        = No
    SM3         = No

SIMD: 256-bit
  * AVX             = Yes
    FMA4            = No
    XOP             = No
  * FMA3            = Yes
  * AVX2            = Yes
    VAES            = Yes
    VPCLMULQDQ      = Yes
  * AVX-GFNI        = No
    AVX-VNNI        = No
    AVX-VNNI_INT8   = No
    AVX-IFMA        = No
    AVX-NE-CONVERT  = No
    AVX-VNNI_INT16  = No
    SHA512          = No
    SM4             = No

SIMD: 512-bit
  * AVX512-F                = No
    AVX512-CD               = No
    AVX512-PF               = No
    AVX512-ER               = No
  * AVX512-VL               = No
  * AVX512-BW               = No
  * AVX512-DQ               = No
    AVX512-VPOPCNTDQ        = No
    AVX512-4VNNIW           = No
    AVX512-4FMAPS           = No
  * AVX512-IFMA             = No
  * AVX512-VBMI             = No
    AVX512-VNNI             = No
  * AVX512-VBMI2            = No
    AVX512-BITALG           = No
    AVX512-VPCLMULQDQ       = No
    AVX512-VAES             = No
  * AVX512-GFNI             = No
    AVX512-BF16             = No
    AVX512-VP2INTERSECT     = No
    AVX512-FP16             = No

APX + AVX10:
    APX-F           = No
    AVX10           = No
    AVX10/128       = No
    AVX10/256       = No
    AVX10/512       = No

AMX:
    AMX-TILE        = No
    AMX-INT8        = No
    AMX-BF16        = No
    AMX-FP16        = No
    AMX-COMPLEX     = No


Auto-Selecting: 19-ZN2 ~ Kagari

/home/fredrik/Downloads/y-cruncher v0.8.7.9547-static/Binaries/19-ZN2 ~ Kagari


Launching y-cruncher...
================================================================

Running from console...


Insufficient permissions to set thread priority. Please retry as root.

Further messages for this warning will be suppressed.

Checking processor/OS features...

Required Features:
    x64, ABM, BMI1, BMI2, ADX,
    SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2,
    AVX, FMA3, AVX2



Reading Hardware Topology...

Logical Cores:
    0-15

Physical Cores:
    0-7

NUMA Node 0:
    0-15

Scheduling Group 0:
    0-15


y-cruncher v0.8.7 Build 9547-icx            ( www.numberworld.org )
Copyright 2008-2025 Alexander J. Yee      ( a-yee@u.northwestern.edu )

Distribute Freely - Please report any bugs.

Binary:  Linux/19-ZN2 ~ Kagari  -  Zen 2 Matisse (x64 ADX)

  0         Benchmark Pi      (standard, memory-intensive)
  1         Benchmark Pi BBP  (digit extraction, CPU only)

  2         Component Stress Tester
  3         Run I/O Performance Analysis

  4         Custom Compute a Constant
                -   Compute other constants (e, Golden Ratio, etc...)
                -   Choose your own settings (use disk for large computations)

  5         BBP Digit Extractor for Pi (Customizer)
  6         Digit Viewer

  7/8       Advanced Options / About
  9         Tuning Profile:  Kagari (20241028)
Enter your choice:
option: 0

Benchmark Pi:

Select a Benchmark Type:

  0     Single-Threaded
  1     Multi-Threaded (all cores)
  2     Multi-Threaded (no SMT)

option: 0


Available Memory:   22.5 GiB

Option      Decimal Digits          Approx. Memory Needed

  1             25,000,000                 125 MiB
  2             50,000,000                 235 MiB
  3            100,000,000                 455 MiB
  4            250,000,000                1.09 GiB
  5            500,000,000                2.17 GiB
  6          1,000,000,000                4.30 GiB
  7          2,500,000,000                10.8 GiB
  8          5,000,000,000                22.3 GiB
  9         10,000,000,000                44.6 GiB
 10         25,000,000,000                 111 GiB
 11         50,000,000,000                 222 GiB
 12        100,000,000,000                 445 GiB
 13        250,000,000,000                1.09 TiB
 14        500,000,000,000                2.18 TiB
 15      1,000,000,000,000                4.48 TiB
 16      2,500,000,000,000                11.2 TiB

  0     I prefer SuperPi sizes... (1M, 2M, 4M...)

option: 3

Constant:   Pi
Algorithm:  Chudnovsky (1988)

Decimal Digits:       100,000,000
Hexadecimal Digits:   83,048,203 (Disabled)

Computation Mode:     Ram Only
Multi-Threading:      None (No Multi-threading)  ->  1 / ?

Start Time: Thu Dec 11 00:01:20 2025

Working Memory...   442 MiB  (locked, spread: ?)                               
Twiddle Tables...  13.9 MiB  (locked, spread: ?)                               

Begin Computation:

Series CommonP2B3...  7,051,378 terms  (Expansion Factor = 2.488)
Time:    16.630 seconds  ( 0.277 minutes )                                     
Large Division...
Time:    1.047 seconds  ( 0.017 minutes )                                      
InvSqrt(10005)...
Time:    0.651 seconds  ( 0.011 minutes )                                      
Large Multiply...
Time:    0.498 seconds  ( 0.008 minutes )                                      

Pi:  18.825 seconds  ( 0.314 minutes )

Processing Hexadecimal Digits:
Time:    0.013 seconds  ( 0.000 minutes )                                      
Base Converting:
Time:    1.562 seconds  ( 0.026 minutes )                                      
Writing Decimal Digits:
Time:    0.090 seconds  ( 0.002 minutes )                                      

Start Time: Thu Dec 11 00:01:20 2025
End Time:   Thu Dec 11 00:01:41 2025

Total Computation Time:     20.388 seconds  ( 0.340 minutes )
Start-to-End Wall Time:     20.793 seconds  ( 0.347 minutes )

CPU Utilization:          100.01 %  +    0.00 % kernel overhead
Multi-core Efficiency:      6.25 %  +    0.00 % kernel overhead

Last Decimal Digits:  Pi
9948682556 3967530560 3352869667 7734610718 4471868529  :  99,999,950
7572203175 2074898161 1683139375 1497058112 0187751592  :  100,000,000

Spot Check: Good through 100,000,000

Program:              y-cruncher v0.8.7 Build 9547-icx
Tuning Profile:       Linux/19-ZN2 ~ Kagari => Kagari (20241028)
Processor(s):         AMD Ryzen 7 PRO 5850U with Radeon Graphics 
Topology:             16 threads / 8 cores / 1 socket / 1 NUMA node
Usable Memory:        32,922,714,112 (30.7 GiB)
Est. CPU Frequency:   4,422,335,609 Hz

Validation File: Pi - 20251211-000141.txt




$ ./y-cruncher 
y-cruncher v0.8.7 Build 9547-gcc

Detecting Environment...

Hardware Features:
(*) Indicates it is used explicitly by y-cruncher.

CPU Vendor:
    AMD         = Yes
    Intel       = No

OS Features:
  * 64-bit      = Yes
  * OS AVX      = Yes
  * OS AVX512   = No

Hardware Features:
    MMX         = Yes
  * x64         = Yes
  * ABM         = Yes
    RDRAND      = Yes
    RDSEED      = Yes
  * BMI1        = Yes
  * BMI2        = Yes
  * ADX         = Yes
    PREFETCHW   = Yes
    PREFETCHWT1 = No
    RDPID       = Yes
    SERIALIZE   = No
    CLDEMOTE    = No

SIMD: 128-bit
  * SSE         = Yes
  * SSE2        = Yes
  * SSE3        = Yes
  * SSSE3       = Yes
    SSE4a       = Yes
  * SSE4.1      = Yes
  * SSE4.2      = Yes
    PCLMULQDQ   = Yes
    AES         = Yes
    SHA         = Yes
    GFNI        = No
    SM3         = No

SIMD: 256-bit
  * AVX             = Yes
    FMA4            = No
    XOP             = No
  * FMA3            = Yes
  * AVX2            = Yes
    VAES            = Yes
    VPCLMULQDQ      = Yes
  * AVX-GFNI        = No
    AVX-VNNI        = No
    AVX-VNNI_INT8   = No
    AVX-IFMA        = No
    AVX-NE-CONVERT  = No
    AVX-VNNI_INT16  = No
    SHA512          = No
    SM4             = No

SIMD: 512-bit
  * AVX512-F                = No
    AVX512-CD               = No
    AVX512-PF               = No
    AVX512-ER               = No
  * AVX512-VL               = No
  * AVX512-BW               = No
  * AVX512-DQ               = No
    AVX512-VPOPCNTDQ        = No
    AVX512-4VNNIW           = No
    AVX512-4FMAPS           = No
  * AVX512-IFMA             = No
  * AVX512-VBMI             = No
    AVX512-VNNI             = No
  * AVX512-VBMI2            = No
    AVX512-BITALG           = No
    AVX512-VPCLMULQDQ       = No
    AVX512-VAES             = No
  * AVX512-GFNI             = No
    AVX512-BF16             = No
    AVX512-VP2INTERSECT     = No
    AVX512-FP16             = No

APX + AVX10:
    APX-F           = No
    AVX10           = No
    AVX10/128       = No
    AVX10/256       = No
    AVX10/512       = No

AMX:
    AMX-TILE        = No
    AMX-INT8        = No
    AMX-BF16        = No
    AMX-FP16        = No
    AMX-COMPLEX     = No


Auto-Selecting: 19-ZN2 ~ Kagari

/home/fredrik/Downloads/y-cruncher v0.8.7.9547-static/Binaries/19-ZN2 ~ Kagari


Launching y-cruncher...
================================================================

Running from console...


Insufficient permissions to set thread priority. Please retry as root.

Further messages for this warning will be suppressed.

Checking processor/OS features...

Required Features:
    x64, ABM, BMI1, BMI2, ADX,
    SSE, SSE2, SSE3, SSSE3, SSE4.1, SSE4.2,
    AVX, FMA3, AVX2



Reading Hardware Topology...

Logical Cores:
    0-15

Physical Cores:
    0-7

NUMA Node 0:
    0-15

Scheduling Group 0:
    0-15


y-cruncher v0.8.7 Build 9547-icx            ( www.numberworld.org )
Copyright 2008-2025 Alexander J. Yee      ( a-yee@u.northwestern.edu )

Distribute Freely - Please report any bugs.

Binary:  Linux/19-ZN2 ~ Kagari  -  Zen 2 Matisse (x64 ADX)

  0         Benchmark Pi      (standard, memory-intensive)
  1         Benchmark Pi BBP  (digit extraction, CPU only)

  2         Component Stress Tester
  3         Run I/O Performance Analysis

  4         Custom Compute a Constant
                -   Compute other constants (e, Golden Ratio, etc...)
                -   Choose your own settings (use disk for large computations)

  5         BBP Digit Extractor for Pi (Customizer)
  6         Digit Viewer

  7/8       Advanced Options / About
  9         Tuning Profile:  Kagari (20241028)
Enter your choice:
option: 0

Benchmark Pi:

Select a Benchmark Type:

  0     Single-Threaded
  1     Multi-Threaded (all cores)
  2     Multi-Threaded (no SMT)

option: 1


Available Memory:   22.5 GiB

Option      Decimal Digits          Approx. Memory Needed

  1             25,000,000                 152 MiB
  2             50,000,000                 269 MiB
  3            100,000,000                 497 MiB
  4            250,000,000                1.12 GiB
  5            500,000,000                2.20 GiB
  6          1,000,000,000                4.56 GiB
  7          2,500,000,000                12.1 GiB
  8          5,000,000,000                22.7 GiB
  9         10,000,000,000                44.7 GiB
 10         25,000,000,000                 111 GiB
 11         50,000,000,000                 222 GiB
 12        100,000,000,000                 445 GiB
 13        250,000,000,000                1.09 TiB
 14        500,000,000,000                2.18 TiB
 15      1,000,000,000,000                4.48 TiB
 16      2,500,000,000,000                11.2 TiB

  0     I prefer SuperPi sizes... (1M, 2M, 4M...)

option: 3

Constant:   Pi
Algorithm:  Chudnovsky (1988)

Decimal Digits:       100,000,000
Hexadecimal Digits:   83,048,203 (Disabled)

Computation Mode:     Ram Only
Multi-Threading:      Push Pool  ->  16 / ? / 16  (all cores)

Start Time: Thu Dec 11 00:04:50 2025

Working Memory...   484 MiB  (locked, spread: ?)                               
Twiddle Tables...  13.9 MiB  (locked, spread: ?)                               

Begin Computation:

Series CommonP2B3...  7,051,378 terms  (Expansion Factor = 2.488)
Time:    3.477 seconds  ( 0.058 minutes )                                      
Large Division...
Time:    0.302 seconds  ( 0.005 minutes )                                      
InvSqrt(10005)...
Time:    0.193 seconds  ( 0.003 minutes )                                      
Large Multiply...
Time:    0.124 seconds  ( 0.002 minutes )                                      

Pi:  4.096 seconds  ( 0.068 minutes )

Processing Hexadecimal Digits:
Time:    0.003 seconds  ( 0.000 minutes )                                      
Base Converting:
Time:    0.326 seconds  ( 0.005 minutes )                                      
Writing Decimal Digits:
Time:    0.064 seconds  ( 0.001 minutes )                                      

Start Time: Thu Dec 11 00:04:50 2025
End Time:   Thu Dec 11 00:04:54 2025

Total Computation Time:     4.422 seconds  ( 0.074 minutes )
Start-to-End Wall Time:     4.800 seconds  ( 0.080 minutes )

CPU Utilization:         1327.18 %  +    7.01 % kernel overhead
Multi-core Efficiency:     82.95 %  +    0.44 % kernel overhead

Last Decimal Digits:  Pi
9948682556 3967530560 3352869667 7734610718 4471868529  :  99,999,950
7572203175 2074898161 1683139375 1497058112 0187751592  :  100,000,000

Spot Check: Good through 100,000,000

Program:              y-cruncher v0.8.7 Build 9547-icx
Tuning Profile:       Linux/19-ZN2 ~ Kagari => Kagari (20241028)
Processor(s):         AMD Ryzen 7 PRO 5850U with Radeon Graphics 
Topology:             16 threads / 8 cores / 1 socket / 1 NUMA node
Usable Memory:        32,922,714,112 (30.7 GiB)
Est. CPU Frequency:   4,409,171,921 Hz

Validation File: Pi - 20251211-000455.txt

[TuomLarsen]

Now that's embarrassing. I didn't know I could run y-cruncher in the interactive mode, there it was clear I was still running with multithreading (-noSMT does something else).

With multithreading disabled I get ~7s for 100M.

The order of universe is restored and I truly apologise for the false alarm.

[fredrik-johansson]

No worries!