From 92ef115fc11689d86a1f5e00e527b2057a4186de Mon Sep 17 00:00:00 2001 From: Marcus Fedarko Date: Tue, 10 Oct 2023 20:49:56 -0700 Subject: [PATCH] DOC/PERF: more 100 Mbp notes --- docs/Benchmarking.ipynb | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/docs/Benchmarking.ipynb b/docs/Benchmarking.ipynb index 558f927..3e3705f 100644 --- a/docs/Benchmarking.ipynb +++ b/docs/Benchmarking.ipynb @@ -1697,7 +1697,7 @@ }, { "cell_type": "markdown", - "id": "f2df2b6d", + "id": "b37eea3c", "metadata": {}, "source": [ "^^^ Note that this test sometimes crashes Jupyter (I got it to run once by closing most of the other tabs / processes I had running on my computer, then rerunning it). I think it's because $k = 5$ is an extremely small value of $k$ to use when we have sequences this big.\n", @@ -1794,7 +1794,7 @@ }, { "cell_type": "markdown", - "id": "ef84854e", + "id": "b6d5adde", "metadata": {}, "source": [ "Yep, this poor computer can handle that much easier." @@ -2012,7 +2012,7 @@ }, { "cell_type": "markdown", - "id": "2709272d", + "id": "20c420a9", "metadata": {}, "source": [ "Yep, this is much easier for wotplot to handle. And the plots might actually be informative!" @@ -2412,9 +2412,9 @@ "id": "ced5cb89", "metadata": {}, "source": [ - "So we can actually create a dot plot of two 100 Mbp sequences (using a large enough $k$) -- it just takes ~50 minutes, at least on my laptop. (The left plot took ~47.4 minutes, and the right plot took ~44.6 minutes.)\n", + "So we can actually create a dot plot of two 100 Mbp sequences (using a large enough $k$) -- it just takes ~50 minutes, at least on my laptop. (The left plot took ~47.4 minutes, and the right plot took ~44.6 minutes. Sometimes you've just gotta... start something running, and then walk over to the gym, and then come back later in the night.)\n", "\n", - "I should probably note here that this result is surprising -- I reran this after adding in support to `viz_spy()` for drawing not-binary matrices in color, thinking that this would cause an increase in total runtime. It's weird that the right plot finished faster than the left plot, even though the actual visualization step took longer! Of course, this might be due to weird variations in other things my laptop was running in the background -- this isn't a very formal benchmarking result." + "I should probably note here that this result is surprising to me -- I reran this after adding in support to `viz_spy()` for drawing not-binary matrices in color, thinking that this would cause an increase in total runtime. It's weird that the right plot finished faster than the left plot, even though the actual visualization step took longer! Of course, this might be due to weird variations in other things my laptop was running in the background -- this isn't a very formal benchmarking result." ] } ],