spmvtest5(1)

✨ spmvtest5 — SpMV kernel benchmark for specified matrix storage format ✨

okay bestie, gather round 💖 this one lives in the Utility Commands part of the manual, and it's literally just here to make your sparse matrix-vector multiply go brrr and tell you how fast it went. no cap.

---

🌸 SYNOPSIS 🌸

spmvtest5 matrix_filename matrix_type iter [block]

that's it, that's the whole vibe. four little slots and you're off 🦋

---

💅 DESCRIPTION 💅

so here's the tea ☕ this program takes the matrix data from matrix_filename, and then it computes the multiply of the matrix with matrix_type and a vector (1, ..., 1)^T. iconic vector choice honestly, all ones, no notes.

the FLOPS performance gets measured as the average of iter iterations. so like, it doesn't just do it once and judge, it does it iter times and averages, which is the correct, main-character way to benchmark ✨

matrix_filename supports the Matrix Market format — that's the one, that's the format, iykyk.

and if you need it, the block size of the BSR and BSC formats can be specified by block. optional queen behavior, only when necessary 💅

for more info, go peek at the Lis User Guide — it has the whole lore.

---

🎀 EXIT STATUS 🎀

here's what you get back when it's done:

• 0 — the process is normally terminated. slay, it worked, we love to see it 🫶
• unspecified — an error occurred 😭 something went wrong and we are simply not going to elaborate further

---

🌈 SEE ALSO 🌈

go read these too, they're all part of the same friend group:

• lis(3)
• lsolve(1)
• esolve(1)
• hpcg_kernel(1)
• hpcg_spmvtest(1)
• spmvtest1(1)
• spmvtest2(1)
• spmvtest2b(1)
• spmvtest3(1)
• spmvtest3b(1)
• spmvtest4(1)

and the links, because of course 💖

• http://www.ssisc.org/lis/
• http://math.nist.gov/MatrixMarket/

---

fr fr, that's the whole page. go benchmark something and come back with numbers 👑

and that's a wrap bestie 🎀 this was pulled straight from the Utility Commands manual page, dated 26 Mar 2014. periodt. 💜