嘟嘟社区

用yabs跑出来的geekbench5分数 普遍比官网的记录高10--15%


用yabs跑出来的geekbench5分数  普遍比官网的记录高10–15%
虚拟化会损失一些性能,还可能超售
普遍高10–15%?你能不能用数据说话或者用代码说话?
https://github.com/masonr/yet-another-bench-script/blob/master/yabs.sh#L756
  1.         if [[ $VERSION == *5* ]]; then # Geekbench v5
  2.                 if [[ $ARCH = *x86* && $GEEKBENCH_4 == *False* ]]; then # don’t run Geekbench 5 if on 32-bit arch
  3.                         echo -e "nGeekbench 5 cannot run on 32-bit architectures. Re-run with -4 flag to use"
  4.                         echo -e "Geekbench 4, which can support 32-bit architectures. Skipping Geekbench 5."
  5.                 elif [[ $ARCH = *x86* && $GEEKBENCH_4 == *True* ]]; then
  6.                         echo -e "nGeekbench 5 cannot run on 32-bit architectures. Skipping test."
  7.                 else
  8.                         echo -en "nRunning GB5 benchmark test… *cue elevator music*"
  9.                         # download the latest Geekbench 5 tarball and extract to geekbench temp directory
  10.                         if [[ $ARCH = *aarch64* || $ARCH = *arm* ]]; then
  11.                                 $DL_CMD https://cdn.geekbench.com/Geekbench-5.4.4-LinuxARMPreview.tar.gz  | tar xz –strip-components=1 -C $GEEKBENCH_PATH &>/dev/null
  12.                         else
  13.                                 $DL_CMD https://cdn.geekbench.com/Geekbench-5.4.4-Linux.tar.gz | tar xz –strip-components=1 -C $GEEKBENCH_PATH &>/dev/null
  14.                         fi
  15.                         # check if geekbench file exists
  16.                         if test -f "geekbench.license"; then
  17.                                 $GEEKBENCH_PATH/geekbench5 –unlock `cat geekbench.license` > /dev/null 2>&1
  18.                         fi
  19.                         GEEKBENCH_TEST=$($GEEKBENCH_PATH/geekbench5 –upload 2>/dev/null | grep "https://browser")
  20.                 fi
  21.         fi
  22.         # ensure the test ran successfully
  23.         if [ -z "$GEEKBENCH_TEST" ]; then
  24.                 if [[ -z "$IPV4_CHECK" ]]; then
  25.                         # Geekbench test failed to download because host lacks IPv4 (cdn.geekbench.com = IPv4 only)
  26.                         echo -e "r 33[0KGeekbench releases can only be downloaded over IPv4. FTP the Geekbench files and run manually."
  27.                 elif [[ $ARCH != *x86* ]]; then
  28.                         # if the Geekbench test failed for any reason, exit cleanly and print error message
  29.                         echo -e "r 33[0KGeekbench $VERSION test failed. Run manually to determine cause."
  30.                 fi
  31.         else
  32.                 # if the Geekbench test succeeded, parse the test results URL
  33.                 GEEKBENCH_URL=$(echo -e $GEEKBENCH_TEST | head -1)
  34.                 GEEKBENCH_URL_CLAIM=$(echo $GEEKBENCH_URL | awk ‘{ print $2 }’)
  35.                 GEEKBENCH_URL=$(echo $GEEKBENCH_URL | awk ‘{ print $1 }’)
  36.                 # sleep a bit to wait for results to be made available on the geekbench website
  37.                 sleep 20
  38.                 # parse the public results page for the single and multi core geekbench scores
  39.                 [[ $VERSION == *5* ]] && GEEKBENCH_SCORES=$($DL_CMD $GEEKBENCH_URL | grep "div class=’score’") ||
  40.                         GEEKBENCH_SCORES=$($DL_CMD $GEEKBENCH_URL | grep "span class=’score’")
  41.                 GEEKBENCH_SCORES_SINGLE=$(echo $GEEKBENCH_SCORES | awk -v FS="(>|<)" ‘{ print $3 }’)
  42.                 GEEKBENCH_SCORES_MULTI=$(echo $GEEKBENCH_SCORES | awk -v FS="(>|<)" ‘{ print $7 }’)
  43.        
  44.                 # print the Geekbench results
  45.                 echo -en "r 33[0K"
  46.                 echo -e "Geekbench $VERSION Benchmark Test:"
  47.                 echo -e "———————————"
  48.                 printf "%-15s | %-30sn" "Test" "Value"
  49.                 printf "%-15s | %-30sn"
  50.                 printf "%-15s | %-30sn" "Single Core" "$GEEKBENCH_SCORES_SINGLE"
  51.                 printf "%-15s | %-30sn" "Multi Core" "$GEEKBENCH_SCORES_MULTI"
  52.                 printf "%-15s | %-30sn" "Full Test" "$GEEKBENCH_URL"
  53.                 # write the geekbench claim URL to a file so the user can add the results to their profile (if desired)
  54.                 [ ! -z "$GEEKBENCH_URL_CLAIM" ] && echo -e "$GEEKBENCH_URL_CLAIM" >> geekbench_claim.url 2> /dev/null
  55.         fi
  56. }
  57. # if the skip geekbench flag was set, skip the system performance test, otherwise test system performance
  58. if [ -z "$SKIP_GEEKBENCH" ]; then
  59.         if [[ $GEEKBENCH_4 == *True* ]]; then
  60.                 launch_geekbench 4
  61.         fi
  62.         if [[ $GEEKBENCH_5 == *True* ]]; then
  63.                 launch_geekbench 5
  64.         fi
  65. fi

复制代码

表妹 发表于 2022-7-21 20:48
普遍高10–15%?你能不能用数据说话或者用代码说话?
https://github.com/masonr/yet-another-bench-script …

贴个别人的代码 有个卵意义啊。。。

我测了6台不同cpu都高出不少,才来说这个现象, 你贴个别人的代码就用事实说话了??

粪红的智商都这样了吗?

nmdwsm 发表于 2022-7-21 20:41
虚拟化会损失一些性能,还可能超售

测了hetzner和ovh的机器,基本都是高不少。

不知道怎么回事。 感觉有点奇怪。