ClickHouse · alexey-milovidov · May 8, 2026 · May 8, 2026 · May 15, 2026
diff --git a/hyrise/Dockerfile b/hyrise/Dockerfile
@@ -0,0 +1,82 @@
+FROM ubuntu:25.04 AS build
+
+ENV DEBIAN_FRONTEND=noninteractive
+ENV HYRISE_HEADLESS_SETUP=1
+
+RUN apt-get update \
+    && apt-get install -y --no-install-recommends \
+        autoconf \
+        bash-completion \
+        bc \
+        ca-certificates \
+        clang-19 \
+        clang-20 \
+        cmake \
+        curl \
+        dos2unix \
+        g++-13 \
+        g++-15 \
+        gcc-13 \
+        gcc-15 \
+        git \
+        libboost-all-dev \
+        libhwloc-dev \
+        libncurses-dev \
+        libnuma-dev \
+        libnuma1 \
+        libpq-dev \
+        libreadline-dev \
+        libsqlite3-dev \
+        libtbb-dev \
+        lld-20 \
+        llvm-20 \
+        lsb-release \
+        make \
+        ninja-build \
+        parallel \
+        postgresql-server-dev-all \
+        python3 \
+        python3-pip \
+        software-properties-common \
+        sudo \
+    && update-alternatives --install /usr/bin/gcc gcc /usr/bin/gcc-15 90 --slave /usr/bin/g++ g++ /usr/bin/g++-15 \
+    && apt-get clean \
+    && rm -rf /var/lib/apt/lists/*
+
+# Pin a specific Hyrise revision for reproducibility (defaults to master).
+ARG HYRISE_REF=master
+# Set NO_LTO=TRUE to disable Link Time Optimization (much faster build, slightly
+# lower runtime performance — useful for development/testing). Default: LTO on.
+ARG NO_LTO=FALSE
+
+WORKDIR /opt
+RUN git clone https://github.com/hyrise/hyrise.git \
+    && cd hyrise \
+    && git checkout "${HYRISE_REF}" \
+    && git submodule update --jobs 8 --init --recursive --depth 1
+
+WORKDIR /opt/hyrise/cmake-build-release
+RUN cmake -GNinja -DCMAKE_BUILD_TYPE=Release -DNO_LTO=${NO_LTO} .. \
+    && ninja hyriseServer
+
+FROM ubuntu:25.04
+ENV DEBIAN_FRONTEND=noninteractive
+RUN apt-get update \
+    && apt-get install -y --no-install-recommends \
+        libboost-system1.83.0 \
+        libboost-thread1.83.0 \
+        libhwloc15 \
+        libnuma1 \
+        libsqlite3-0 \
+        libtbb12 \
+    && apt-get clean \
+    && rm -rf /var/lib/apt/lists/*
+
+COPY --from=build /opt/hyrise/cmake-build-release/hyriseServer /usr/local/bin/hyriseServer
+COPY --from=build /opt/hyrise/cmake-build-release/lib/libhyrise_impl.so /usr/local/lib/
+COPY --from=build /opt/hyrise/cmake-build-release/third_party/jemalloc/lib/libjemalloc.so.2 /usr/local/lib/
+RUN ldconfig
+
+EXPOSE 5432
+ENTRYPOINT ["/usr/local/bin/hyriseServer"]
+CMD ["5432"]
diff --git a/hyrise/benchmark.sh b/hyrise/benchmark.sh
@@ -0,0 +1,11 @@
+#!/bin/bash
+# Thin shim — actual flow is in lib/benchmark-common.sh.
+#
+# Hyrise has no on-disk persistence: every restart resurfaces with an empty
+# catalog and the dataset must be re-loaded into RAM. BENCH_DURABLE=no makes
+# the driver re-run ./load on every cold cycle (and roll that wall-clock into
+# the cold-try timing) so each "cold" number genuinely measures
+# load+query against a fresh in-memory dataset.
+export BENCH_DOWNLOAD_SCRIPT="download-hits-csv"
+export BENCH_DURABLE=no
+exec ../lib/benchmark-common.sh
diff --git a/hyrise/check b/hyrise/check
@@ -0,0 +1,4 @@
+#!/bin/bash
+set -e
+
+psql -h 127.0.0.1 -p 5432 -U postgres -c 'SELECT 1' >/dev/null
diff --git a/hyrise/data-size b/hyrise/data-size
@@ -0,0 +1,10 @@
+#!/bin/bash
+set -eu
+
+# Hyrise has no on-disk persistence; report the total in-memory segment
+# size from meta_segments. Strip any whitespace so the driver's strict
+# digits-only regex check passes.
+psql -h 127.0.0.1 -p 5432 -U postgres -t -A \
+    -c "SELECT SUM(estimated_size_in_bytes) FROM meta_segments WHERE table_name = 'hits';" \
+    | tr -d '[:space:]'
+echo
diff --git a/hyrise/hits.csv.json b/hyrise/hits.csv.json
@@ -0,0 +1,116 @@
+{
+    "config": {
+        "rfc_mode": true,
+        "separator": ",",
+        "quote": "\"",
+        "escape": "\"",
+        "delimiter": "\n"
+    },
+    "columns": [
+        {"name": "WatchID",               "type": "long",   "nullable": false},
+        {"name": "JavaEnable",            "type": "int",    "nullable": false},
+        {"name": "Title",                 "type": "string", "nullable": false},
+        {"name": "GoodEvent",             "type": "int",    "nullable": false},
+        {"name": "EventTime",             "type": "string", "nullable": false},
+        {"name": "EventDate",             "type": "string", "nullable": false},
+        {"name": "CounterID",             "type": "int",    "nullable": false},
+        {"name": "ClientIP",              "type": "int",    "nullable": false},
+        {"name": "RegionID",              "type": "int",    "nullable": false},
+        {"name": "UserID",                "type": "long",   "nullable": false},
+        {"name": "CounterClass",          "type": "int",    "nullable": false},
+        {"name": "OS",                    "type": "int",    "nullable": false},
+        {"name": "UserAgent",             "type": "int",    "nullable": false},
+        {"name": "URL",                   "type": "string", "nullable": false},
+        {"name": "Referer",               "type": "string", "nullable": false},
+        {"name": "IsRefresh",             "type": "int",    "nullable": false},
+        {"name": "RefererCategoryID",     "type": "int",    "nullable": false},
+        {"name": "RefererRegionID",       "type": "int",    "nullable": false},
+        {"name": "URLCategoryID",         "type": "int",    "nullable": false},
+        {"name": "URLRegionID",           "type": "int",    "nullable": false},
+        {"name": "ResolutionWidth",       "type": "int",    "nullable": false},
+        {"name": "ResolutionHeight",      "type": "int",    "nullable": false},
+        {"name": "ResolutionDepth",       "type": "int",    "nullable": false},
+        {"name": "FlashMajor",            "type": "int",    "nullable": false},
+        {"name": "FlashMinor",            "type": "int",    "nullable": false},
+        {"name": "FlashMinor2",           "type": "string", "nullable": false},
+        {"name": "NetMajor",              "type": "int",    "nullable": false},
+        {"name": "NetMinor",              "type": "int",    "nullable": false},
+        {"name": "UserAgentMajor",        "type": "int",    "nullable": false},
+        {"name": "UserAgentMinor",        "type": "string", "nullable": false},
+        {"name": "CookieEnable",          "type": "int",    "nullable": false},
+        {"name": "JavascriptEnable",      "type": "int",    "nullable": false},
+        {"name": "IsMobile",              "type": "int",    "nullable": false},
+        {"name": "MobilePhone",           "type": "int",    "nullable": false},
+        {"name": "MobilePhoneModel",      "type": "string", "nullable": false},
+        {"name": "Params",                "type": "string", "nullable": false},
+        {"name": "IPNetworkID",           "type": "int",    "nullable": false},
+        {"name": "TraficSourceID",        "type": "int",    "nullable": false},
+        {"name": "SearchEngineID",        "type": "int",    "nullable": false},
+        {"name": "SearchPhrase",          "type": "string", "nullable": false},
+        {"name": "AdvEngineID",           "type": "int",    "nullable": false},
+        {"name": "IsArtifical",           "type": "int",    "nullable": false},
+        {"name": "WindowClientWidth",     "type": "int",    "nullable": false},
+        {"name": "WindowClientHeight",    "type": "int",    "nullable": false},
+        {"name": "ClientTimeZone",        "type": "int",    "nullable": false},
+        {"name": "ClientEventTime",       "type": "string", "nullable": false},
+        {"name": "SilverlightVersion1",   "type": "int",    "nullable": false},
+        {"name": "SilverlightVersion2",   "type": "int",    "nullable": false},
+        {"name": "SilverlightVersion3",   "type": "int",    "nullable": false},
+        {"name": "SilverlightVersion4",   "type": "int",    "nullable": false},
+        {"name": "PageCharset",           "type": "string", "nullable": false},
+        {"name": "CodeVersion",           "type": "int",    "nullable": false},
+        {"name": "IsLink",                "type": "int",    "nullable": false},
+        {"name": "IsDownload",            "type": "int",    "nullable": false},
+        {"name": "IsNotBounce",           "type": "int",    "nullable": false},
+        {"name": "FUniqID",               "type": "long",   "nullable": false},
+        {"name": "OriginalURL",           "type": "string", "nullable": false},
+        {"name": "HID",                   "type": "int",    "nullable": false},
+        {"name": "IsOldCounter",          "type": "int",    "nullable": false},
+        {"name": "IsEvent",               "type": "int",    "nullable": false},
+        {"name": "IsParameter",           "type": "int",    "nullable": false},
+        {"name": "DontCountHits",         "type": "int",    "nullable": false},
+        {"name": "WithHash",              "type": "int",    "nullable": false},
+        {"name": "HitColor",              "type": "string", "nullable": false},
+        {"name": "LocalEventTime",        "type": "string", "nullable": false},
+        {"name": "Age",                   "type": "int",    "nullable": false},
+        {"name": "Sex",                   "type": "int",    "nullable": false},
+        {"name": "Income",                "type": "int",    "nullable": false},
+        {"name": "Interests",             "type": "int",    "nullable": false},
+        {"name": "Robotness",             "type": "int",    "nullable": false},
+        {"name": "RemoteIP",              "type": "int",    "nullable": false},
+        {"name": "WindowName",            "type": "int",    "nullable": false},
+        {"name": "OpenerName",            "type": "int",    "nullable": false},
+        {"name": "HistoryLength",         "type": "int",    "nullable": false},
+        {"name": "BrowserLanguage",       "type": "string", "nullable": false},
+        {"name": "BrowserCountry",        "type": "string", "nullable": false},
+        {"name": "SocialNetwork",         "type": "string", "nullable": false},
+        {"name": "SocialAction",          "type": "string", "nullable": false},
+        {"name": "HTTPError",             "type": "int",    "nullable": false},
+        {"name": "SendTiming",            "type": "int",    "nullable": false},
+        {"name": "DNSTiming",             "type": "int",    "nullable": false},
+        {"name": "ConnectTiming",         "type": "int",    "nullable": false},
+        {"name": "ResponseStartTiming",   "type": "int",    "nullable": false},
+        {"name": "ResponseEndTiming",     "type": "int",    "nullable": false},
+        {"name": "FetchTiming",           "type": "int",    "nullable": false},
+        {"name": "SocialSourceNetworkID", "type": "int",    "nullable": false},
+        {"name": "SocialSourcePage",      "type": "string", "nullable": false},
+        {"name": "ParamPrice",            "type": "long",   "nullable": false},
+        {"name": "ParamOrderID",          "type": "string", "nullable": false},
+        {"name": "ParamCurrency",         "type": "string", "nullable": false},
+        {"name": "ParamCurrencyID",       "type": "int",    "nullable": false},
+        {"name": "OpenstatServiceName",   "type": "string", "nullable": false},
+        {"name": "OpenstatCampaignID",    "type": "string", "nullable": false},
+        {"name": "OpenstatAdID",          "type": "string", "nullable": false},
+        {"name": "OpenstatSourceID",      "type": "string", "nullable": false},
+        {"name": "UTMSource",             "type": "string", "nullable": false},
+        {"name": "UTMMedium",             "type": "string", "nullable": false},
+        {"name": "UTMCampaign",           "type": "string", "nullable": false},
+        {"name": "UTMContent",            "type": "string", "nullable": false},
+        {"name": "UTMTerm",               "type": "string", "nullable": false},
+        {"name": "FromTag",               "type": "string", "nullable": false},
+        {"name": "HasGCLID",              "type": "int",    "nullable": false},
+        {"name": "RefererHash",           "type": "long",   "nullable": false},
+        {"name": "URLHash",               "type": "long",   "nullable": false},
+        {"name": "CLID",                  "type": "int",    "nullable": false}
+    ]
+}
diff --git a/hyrise/install b/hyrise/install
@@ -0,0 +1,17 @@
+#!/bin/bash
+set -eu
+
+sudo apt-get update -y
+sudo apt-get install -y docker.io postgresql-client gzip
+
+# Hyrise has no upstream binary distribution and the build requires a
+# recent toolchain (gcc-15 / clang-20). Build inside Docker on top of
+# Ubuntu 25.04, then ship just the binary and its runtime libs in a slim
+# image.
+sudo docker build -t clickbench-hyrise .
+
+# The bind-mounted /data dir must exist before docker run (created here
+# instead of in ./load so ./start can be called before the first ./load
+# without erroring on a missing mount source).
+mkdir -p data
+chmod 777 data
diff --git a/hyrise/load b/hyrise/load
@@ -0,0 +1,22 @@
+#!/bin/bash
+set -eu
+
+# Stage hits.csv next to hits.csv.json under data/. The bind mount on
+# the container exposes this dir as /data:ro.
+#
+# Idempotent: BENCH_DURABLE=no triggers ./load again on every cold cycle.
+# The first invocation moves hits.csv (delivered into cwd by
+# download-hits-csv) into data/; subsequent invocations reuse the staged
+# file because the dataset is many GB and re-downloading it per cycle
+# would blow up the run time without changing the measurement.
+if [ -f hits.csv ]; then
+    cp hits.csv.json data/
+    mv -f hits.csv data/hits.csv
+fi
+
+# COPY into a not-yet-existing table works because hits.csv.json next to
+# the data file tells Hyrise's CSV parser the column types — running
+# CREATE TABLE first and then COPY trips an internal Hyrise assertion.
+psql -h 127.0.0.1 -p 5432 -U postgres -v ON_ERROR_STOP=1 -q \
+    -c "COPY hits FROM '/data/hits.csv' WITH (FORMAT CSV);"
+sync