Merge pull request #22 from opensolvers/site-updates

Add Quantum ESPRESSO app page

Author: hugomeiland

README.md

@@ -4,7 +4,7 @@ OpenSolvers explores how open-source scientific software runs on real hardware 
 
 We benchmark **scientific libraries** and **applications** on consumer RISC-V boards through the [EESSI](https://www.eessi.io/) stack — from BLAS kernels up to full app runs — swapping fixed OpenBLAS builds via FlexiBLAS without rebuilding downstream code.
 
-Recent highlights on the Orange Pi RV2 (SpaceMiT X60, RVV): fixing an OpenBLAS `gemv_n` bug restores correctness across [BLAS](scientific-libs/blas.html), [LAPACK](scientific-libs/lapack.html), [ELPA](scientific-libs/elpa.html), and [HPL](apps/hpl.html) — with patched RVV reaching **10.53 GFLOP/s** on Linpack and **1.58×** on a dense eigensolve.
+Recent highlights on the Orange Pi RV2 (SpaceMiT X60, RVV): fixing an OpenBLAS `gemv_n` bug restores correctness across [BLAS](scientific-libs/blas.html), [LAPACK](scientific-libs/lapack.html), [ELPA](scientific-libs/elpa.html), [HPL](apps/hpl.html), and [Quantum ESPRESSO](apps/qe.html) — with patched RVV reaching **10.53 GFLOP/s** on Linpack, **1.58×** on a dense eigensolve, and **1.31×** on a 64-atom Si DFT SCF.
 
 ## Scientific libs
 
@@ -21,6 +21,7 @@ Library-level probes — performance *and* numerical correctness:
 End-to-end application benchmarks on the same boards and EESSI toolchain:
 
 - **[HPL](apps/hpl.html)** — High Performance Linpack; cross-board summary and A/B configs from [opensolvers/benchmarks](https://github.com/opensolvers/benchmarks)
+- **[Quantum ESPRESSO](apps/qe.html)** — plane-wave DFT SCF (`pw.x`); whole-application BLAS backend A/B with per-routine timers
 
 ## Boards
 

_config.yml

@@ -17,6 +17,8 @@ navigation_boards:
 navigation_apps:
   - title: HPL
     url: /apps/hpl.html
+  - title: Quantum ESPRESSO
+    url: /apps/qe.html
 
 navigation_scientific_libs:
   - title: BLAS

apps/qe.md

@@ -0,0 +1,68 @@
+# Quantum ESPRESSO
+
+[Quantum ESPRESSO](https://www.quantum-espresso.org/) `pw.x` plane-wave DFT SCF — a **whole-application** BLAS backend A/B. Swaps only the OpenBLAS implementation (via FlexiBLAS) under one unchanged `pw.x` binary; QE comes from the EESSI module (`QuantumESPRESSO/7.5-foss-2025b`).
+
+Benchmark source: [opensolvers/benchmarks/qe](https://github.com/opensolvers/benchmarks/tree/main/qe) — inputs, a Si supercell generator, and `run-qe-ab.sh` / `run-perf-ab.sh` runners.
+
+## Why QE as a probe
+
+[`dgemm`](../scientific-libs/dgemm.html) isolates BLAS-3; [HPL](hpl.html) and [ELPA](../scientific-libs/elpa.html) probe single HPC solvers. A full DFT SCF mixes level-3 GEMM (`calbec`, subspace rotation), dense LAPACK diagonalization, latency-bound BLAS-2, MPI, **and a large FFT fraction** — so it shows both whether a buggy vector BLAS breaks a production code, and what fraction of a real run the RVV speedup actually moves.
+
+## Correctness — stock RVV `gemv_n` breaks a real DFT SCF
+
+`run-qe-ab.sh si-scf.in 4` on SpaceMiT X60 ([Banana Pi F3](../boards/F3.html)), same `pw.x`, backend swapped via FlexiBLAS:
+
+| Backend | Result | Total energy |
+| ------- | ------ | ------------ |
+| Stock EESSI OpenBLAS 0.3.30, default RVV | `Error in routine inverse_s (1)` → MPI_ABORT | *(none)* |
+| Scalar (`RISCV64_GENERIC`) | converged, 7 iterations | **−14.57861334 Ry** |
+| Patched RVV (`gemv_n` fix) | converged, 7 iterations (identical trace) | **−14.57861334 Ry** |
+
+The stock vector backend aborts in `inverse_s` (overlap-matrix inversion / Lowdin orthonormalization, which leans on `dgemv`) — the same OpenBLAS 0.3.30 RVV `gemv_n` bug that NaNs [HPL](hpl.html) and [ELPA](../scientific-libs/elpa.html). The patched vector build reproduces the scalar SCF bit-for-bit.
+
+## Performance — 64-atom Si supercell
+
+`run-perf-ab.sh si-super-64.in 4` — gamma-point, np=4, `-ndiag 1` (keeps subspace diagonalization on serial LAPACK→OpenBLAS). Measured on [Banana Pi F3](../boards/F3.html) (SpaceMiT X60, 3.7 GB RAM).
+
+### High band count (`nbnd=272`)
+
+| Routine | Scalar | Patched RVV | Speedup | Kind |
+| ------- | -----: | ----------: | ------: | ---- |
+| `calbec` (`<β\|ψ>`) | 6.25 s | 3.18 s | **1.97×** | DGEMM |
+| Subspace rotation (in `regterg`) | 33.2 s | 16.6 s | **2.0×** | DGEMM |
+| `rdiaghg` (dense diag) | 28.2 s | 18.0 s | **1.57×** | LAPACK→BLAS3 |
+| `vloc_psi` (apply V to bands) | 52.6 s | 52.4 s | 1.00× | **FFT (untouched)** |
+| `fftw` | 54.5 s | 53.8 s | 1.01× | **FFT (untouched)** |
+| **`PWSCF` total** | **144.8 s** | **110.3 s** | **1.31×** | whole SCF |
+
+### Default band count (`nbnd=136`)
+
+**67.6 s → 57.0 s = 1.19×** overall (`calbec` 1.90×, `rdiaghg` 1.53×). BLAS routines speed up ~1.5–2.0×; the FFT half of the run does not move (FlexiBLAS only swaps BLAS), which caps the whole-application number.
+
+## Where QE sits on the BLAS-dilution spectrum
+
+Same X60, patched RVV vs scalar:
+
+| Probe | Speedup | Why |
+| ----- | ------: | --- |
+| [DGEMM](../scientific-libs/dgemm.html) (pure level-3) | ~2.3× | all BLAS-3 |
+| [HPL](hpl.html) (Linpack) | ~1.8× | BLAS-3 + `dgemv` panel factorization |
+| [ELPA](../scientific-libs/elpa.html) (eigensolver) | ~1.58× | BLAS-3 + BLAS-2 tridiagonalization |
+| **QE** (full DFT SCF) | **~1.2–1.3×** | BLAS + ~40–50% FFT + MPI |
+
+Each step down adds more non-BLAS / latency-bound work, diluting the BLAS-3 peak. QE is the most realistic whole-application figure — and the reason a raw `dgemm` A/B *overstates* what a real code sees.
+
+## Reproducing
+
+```bash
+module load QuantumESPRESSO/7.5-foss-2025b
+curl -O https://pseudopotentials.quantum-espresso.org/upf_files/Si.pz-vbc.UPF
+
+# Correctness (np=4)
+RVV_LIB=/path/to/patched/libopenblas.so ./run-qe-ab.sh si-scf.in 4
+
+# Performance (np=4)
+RVV_LIB=/path/to/patched/libopenblas.so ./run-perf-ab.sh si-super-64.in 4
+```
+
+**Toolchain:** QuantumESPRESSO 7.5 / foss-2025b (EESSI), FlexiBLAS 3.4.5, OpenBLAS 0.3.30. Patched vector backend = OpenBLAS 0.3.30 with the RISC-V `gemv_n` NaN fix backported — same build used by [DGEMM](../scientific-libs/dgemm.html), [HPL](hpl.html), and [ELPA](../scientific-libs/elpa.html).

boards/F3.md

@@ -33,5 +33,6 @@ See also the [HPL app overview](../apps/hpl.html).
 | [NumPy](../scientific-libs/numpy.html) DGEMM N=4096 | 4.91 GFLOP/s | 17.51 GFLOP/s | 3.6× |
 | [NumPy](../scientific-libs/numpy.html) `eigvalsh` N=2048 | 9.59 s | 5.94 s | 1.6× |
 | ELPA `na=3000` | 50.42 s | **34.83 s** | 1.45× |
+| [QE](../apps/qe.html) `si-super-64.in` (PWSCF, nbnd=272) | 144.8 s | **110.3 s** | 1.31× |
 
-Stock unpatched RVV: HPL and ELPA fail (`nan`); NumPy `eigvalsh` raises `LinAlgError: Eigenvalues did not converge`.
+Stock unpatched RVV: HPL, ELPA, and [Quantum ESPRESSO](../apps/qe.html) fail (`nan` / SCF abort); NumPy `eigvalsh` raises `LinAlgError: Eigenvalues did not converge`.

scientific-libs/dgemm.md

@@ -31,7 +31,7 @@ Bit-identical on [Orange Pi RV2](../boards/RV2.html) and [Banana Pi F3](../board
 | Forced scalar | 0 | 0 | 42.06549 (reference) |
 | Patched RVV (`gemv_n` fix) | 0 | 0 | 42.06549 (matches) |
 
-Fault is in **`dgemv` only** — plain `dgemm` looks fine on the broken build, which is why [HPL](../apps/hpl.html) can NaN while a GEMM micro-benchmark passes.
+Fault is in **`dgemv` only** — plain `dgemm` looks fine on the broken build, which is why [HPL](../apps/hpl.html) and [Quantum ESPRESSO](../apps/qe.html) can fail while a GEMM micro-benchmark passes.
 
 ## Performance — `bench_dgemm`
 

scientific-libs/elpa.md

@@ -14,7 +14,7 @@ Benchmark source: [opensolvers/benchmarks/elpa](https://github.com/opensolvers/b
 | Scalar (`RISCV64_GENERIC`) | 54.92 s | finite=1 |
 | Patched RVV (`gemv_n` fix) | **34.81 s** | finite=1 |
 
-The stock vector backend is faster than scalar but **wrong** — the same OpenBLAS 0.3.30 `gemv_n` NaN bug that breaks [HPL](../apps/hpl.html). With the patch, RVV is **1.58×** faster than scalar (54.92 / 34.81) and numerically correct.
+The stock vector backend is faster than scalar but **wrong** — the same OpenBLAS 0.3.30 `gemv_n` NaN bug that breaks [HPL](../apps/hpl.html) and [Quantum ESPRESSO](../apps/qe.html). With the patch, RVV is **1.58×** faster than scalar (54.92 / 34.81) and numerically correct.
 
 Speedup is smaller than pure `dgemm` (~2.3×) because tridiagonalization is latency-bound BLAS-2 — by design.
 

website.md

@@ -19,7 +19,7 @@ Context file for structural and content decisions on [opensolvers.com](https://w
 
 1. **Home**
 2. **Boards** — VisionFive 2, OrangePi RV2, BananaPi F3
-3. **Apps** — HPL
+3. **Apps** — HPL, Quantum ESPRESSO
 4. **Scientific libs** — BLAS, DGEMM, NumPy, LAPACK, ELPA
 
 Nav config: `_config.yml` (`navigation`, `navigation_boards`, `navigation_apps`, `navigation_scientific_libs`). Rendered in `_includes/header.html`. Cayman theme requires `_layouts/default.html` override to include the header.
@@ -28,7 +28,7 @@ Nav config: `_config.yml` (`navigation`, `navigation_boards`, `navigation_apps`,
 
 | Source | Used for |
 |--------|----------|
-| [opensolvers/benchmarks](https://github.com/opensolvers/benchmarks) | BLAS verification (`dgemm/`), HPL configs (`hpl/`), ELPA (`elpa/`), NumPy/LAPACK (`numpy/`) |
+| [opensolvers/benchmarks](https://github.com/opensolvers/benchmarks) | BLAS verification (`dgemm/`), HPL configs (`hpl/`), ELPA (`elpa/`), QE (`qe/`), NumPy/LAPACK (`numpy/`) |
 | [EESSI/docs#818](https://github.com/EESSI/docs/pull/818) | VisionFive 2 / U74 OpenBLAS + HPL |
 | [EESSI/docs#819](https://github.com/EESSI/docs/pull/819) | Orange Pi RV2 / X60 RVV `gemv_n` fix + HPL |
 | [easyconfigs#26436](https://github.com/easybuilders/easybuild-easyconfigs/pull/26436) | U74 OpenBLAS package |
@@ -44,6 +44,7 @@ Nav config: `_config.yml` (`navigation`, `navigation_boards`, `navigation_apps`,
 
 | Date | Decision |
 |------|----------|
+| 2026-07-14 | Add Quantum ESPRESSO app page from `opensolvers/benchmarks/qe` |
 | 2026-07-14 | IME (X60 `smt.vmadot`) section on RV2 and BPI-F3 board pages |
 | 2026-07-14 | Add dedicated DGEMM and NumPy scientific-lib pages from benchmarks repo |
 | 2026-07-14 | BPI-F3 results from `opensolvers/benchmarks`: HPL 11.52 GFLOP/s, 3.7 GB RAM limit |