i'd guess upgrading your cpu would have more benefit --- but note that
more threads is more important than single-core performance.
On Mon, Sep 18, 2023 at 7:16 PM 洪归风 wrote:
>
> Can you reduce compile times simply by upgrading to faster memory? I'm
> encountering a similar issue where it takes
Can you reduce compile times simply by upgrading to faster memory? I'm
encountering a similar issue where it takes about 10 minutes to generate
the ninja file. I currently have 64GB of memory, but it's running at a
speed of only 2133MHz. Therefore, I'm wondering if I should upgrade my
memory or
I feel you. I splurged on a Threadripper Pro 24-core machine with 128 GB of
RAM and 960 GB Intel optane source code storage drive because I wanted to
spend what I could afford to minimize the time I spend waiting for Android
ROM builds.
On Tue, Mar 14, 2023 at 11:24 AM 'enh' via Android Building <
depends how much you value your time. we have graphs somewhere where we
actually did the experiment of adding memory and cores. iirc the impact of
more memory does top out somewhere around 128GiB, but since everything
grows and 256 is the next step up, that gives some breathing room to keep
these m
You purchased way to much, overkill.
On Tue, Mar 14, 2023 at 10:56 AM 'enh' via Android Building <
android-building@googlegroups.com> wrote:
> yeah, unfortunately i'm only aware of two options, and they're both
> basically "throw money at the problem until it goes away":
>
> 1. get a better compu
yeah, unfortunately i'm only aware of two options, and they're both
basically "throw money at the problem until it goes away":
1. get a better computer. (256 GiB of ram, 64 cores, and 2 TiB of ssd are
the current default.)
2. build in the cloud.
https://cs.android.com/android/platform/superprojec
I overcame that issue, changing 18 minutes into 3 minutes now. By changing
my 24gb of memory that was only 2400mhz., into 32gb of memory at 3200mhz.
And it was unbearable for me also. Soong only analyzes after certain things
you do in the source code. For sure, after a mk or or bp file change. It
w
Any hints from Googlers on how to improve this?
On Saturday, March 4, 2023 at 6:08:04 AM UTC+11 Dean Wheatley wrote:
> Wow, 18 minutes, that's unbearable.
>
> On Tuesday, February 7, 2023 at 7:01:21 AM UTC+11 Michael Swartout wrote:
>
>> Lucky yours only takes 60 seconds. It takes 18 minutes, ever
Wow, 18 minutes, that's unbearable.
On Tuesday, February 7, 2023 at 7:01:21 AM UTC+11 Michael Swartout wrote:
> Lucky yours only takes 60 seconds. It takes 18 minutes, everytime I make a
> change to almost anything.
>
> On Wed, Feb 1, 2023 at 9:29 PM Dean Wheatley wrote:
>
>> Is there a better/
(yeah, anecdotally "about a minute" sounds about right for the
machines developers at google have. i've seen 2-3 minutes too.)
On Mon, Feb 6, 2023 at 12:01 PM Michael Swartout
wrote:
>
> Lucky yours only takes 60 seconds. It takes 18 minutes, everytime I make a
> change to almost anything.
>
> O
Lucky yours only takes 60 seconds. It takes 18 minutes, everytime I make a
change to almost anything.
On Wed, Feb 1, 2023 at 9:29 PM Dean Wheatley wrote:
> Is there a better/faster way to develop Android.bp in the AOSP tree than
> waiting for the many (> 60) seconds it takes to "analyzing Androi
Is there a better/faster way to develop Android.bp in the AOSP tree than
waiting for the many (> 60) seconds it takes to "analyzing Android.bp files
and generating ninja file at out/soong/build.ninja" after any change to any
Android.bp file in the tree?
--
--
You received this message because
12 matches
Mail list logo