Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > alt.comp.os.windows-11 > #18459 > unrolled thread

memory usage vs fan usage

Back to article view | Back to alt.comp.os.windows-11

Contents

  memory usage vs fan usage Andy Burns <usenet@andyburns.uk> - 2025-04-17 15:16 +0100
    Re: memory usage vs fan usage Paul <nospam@needed.invalid> - 2025-04-17 12:09 -0400
      Re: memory usage vs fan usage Andy Burns <usenet@andyburns.uk> - 2025-04-17 19:57 +0100
        Re: memory usage vs fan usage Paul <nospam@needed.invalid> - 2025-04-18 03:15 -0400
    Re: memory usage vs fan usage VanguardLH <V@nguard.LH> - 2025-04-17 11:59 -0500
      Re: memory usage vs fan usage Andy Burns <usenet@andyburns.uk> - 2025-04-17 19:40 +0100
        Re: memory usage vs fan usage Frank Slootweg <this@ddress.is.invalid> - 2025-04-19 12:04 +0000
          Re: memory usage vs fan usage Andy Burns <usenet@andyburns.uk> - 2025-04-19 13:36 +0100
            Re: memory usage vs fan usage Paul <nospam@needed.invalid> - 2025-04-19 09:25 -0400
            Re: memory usage vs fan usage Frank Slootweg <this@ddress.is.invalid> - 2025-04-19 14:54 +0000
              Re: memory usage vs fan usage Andy Burns <usenet@andyburns.uk> - 2025-04-19 16:16 +0100
              Re: memory usage vs fan usage Andy Burns <usenet@andyburns.uk> - 2025-04-24 08:17 +0100
                Re: memory usage vs fan usage VanguardLH <V@nguard.LH> - 2025-04-24 05:14 -0500
                  Re: memory usage vs fan usage Andy Burns <usenet@andyburns.uk> - 2025-04-24 11:30 +0100
                    Re: memory usage vs fan usage Paul <nospam@needed.invalid> - 2025-04-24 09:08 -0400
                    Re: memory usage vs fan usage VanguardLH <V@nguard.LH> - 2025-04-24 20:59 -0500
                      Re: memory usage vs fan usage Andy Burns <usenet@andyburns.uk> - 2025-04-25 07:40 +0100
                        Re: memory usage vs fan usage Andy Burns <usenet@andyburns.uk> - 2025-04-25 19:13 +0100
                          Re: memory usage vs fan usage VanguardLH <V@nguard.LH> - 2025-04-25 22:29 -0500
                            Re: memory usage vs fan usage Andy Burns <usenet@andyburns.uk> - 2025-04-26 10:31 +0100
                              Re: memory usage vs fan usage VanguardLH <V@nguard.LH> - 2025-04-26 15:05 -0500
          Re: memory usage vs fan usage "Carlos E.R." <robin_listas@es.invalid> - 2025-04-19 14:36 +0200
            Re: memory usage vs fan usage Andy Burns <usenet@andyburns.uk> - 2025-04-19 13:56 +0100
            Re: memory usage vs fan usage Paul <nospam@needed.invalid> - 2025-04-19 09:28 -0400
    Re: memory usage vs fan usage VanguardLH <V@nguard.LH> - 2025-04-17 15:07 -0500
      Re: memory usage vs fan usage "Carlos E.R." <robin_listas@es.invalid> - 2025-04-17 22:25 +0200
        Re: memory usage vs fan usage VanguardLH <V@nguard.LH> - 2025-04-17 19:30 -0500
          Re: memory usage vs fan usage "Carlos E.R." <robin_listas@es.invalid> - 2025-04-18 14:09 +0200
            Re: memory usage vs fan usage VanguardLH <V@nguard.LH> - 2025-04-18 11:13 -0500
              Re: memory usage vs fan usage "Carlos E.R." <robin_listas@es.invalid> - 2025-04-18 21:33 +0200
      Re: memory usage vs fan usage Andy Burns <usenet@andyburns.uk> - 2025-04-17 22:12 +0100
        Re: memory usage vs fan usage VanguardLH <V@nguard.LH> - 2025-04-17 20:07 -0500
        Re: memory usage vs fan usage Paul <nospam@needed.invalid> - 2025-04-18 05:24 -0400
          Re: memory usage vs fan usage Andy Burns <usenet@andyburns.uk> - 2025-04-18 12:08 +0100
            Re: memory usage vs fan usage Frank Slootweg <this@ddress.is.invalid> - 2025-04-18 18:09 +0000
              Re: memory usage vs fan usage Andy Burns <usenet@andyburns.uk> - 2025-04-18 19:36 +0100
    Re: memory usage vs fan usage "Carlos E.R." <robin_listas@es.invalid> - 2025-04-17 22:29 +0200
      Re: memory usage vs fan usage Andy Burns <usenet@andyburns.uk> - 2025-04-17 22:17 +0100
        Re: memory usage vs fan usage "Carlos E.R." <robin_listas@es.invalid> - 2025-04-18 02:19 +0200
        Re: memory usage vs fan usage Paul <nospam@needed.invalid> - 2025-04-18 05:52 -0400

Page 2 of 2 — ← Prev page 1 [2]

#18713

Andy Burns <usenet@andyburns.uk> wrote:

> VanguardLH wrote:
> 
>> Just because Microsoft has drivers on their WU
>> server doesn't mean those are the best drivers, or even the correct
>> ones.  What is latest at the WU server may not be what is latest from
>> the hardware maker.
> 
> Intel and Lenovo have a habit of getting into a driver upgrade/downgrade 
> war on both my laptops ... never seemed to happen with Dells, Intel 
> always won ...

Oh, that reminds me: there can be bundleware on laptops for auto-updater
from the computer maker, or separately from the WU scheme.  For example,
Intel has their auto-updater (Intel DSA = Intel Driver & Support) that
could conflict regarding which is latest or best with WU, or other
auto-updaters.  I had Intel DSA on my desktop PC for a short time, but
got rid of it when it suggested drivers that were not for my particular
hardware.  Plus, if the hardware is fully functional, and an updater
addresses issues that you do not experience or to which you will not be
exposed, there is no point in updating.  Don't fix what alread works.
New code may fix some problems, but it can also introduce new problems.

[toc] | [prev] | [next] | [standalone]

#18501

On 2025-04-19 14:04, Frank Slootweg wrote:
>    IIRC, SpeedFan [1] can show temperatures (and fan speed). Don't know
> if it can do both at the same time.

I use a tool that is designed for Linux, but also exists, with 
limitations, for Windows.

https://en.wikipedia.org/wiki/GKrellM

Home page: <https://gkrellm.srcbox.net/>. There is a link for Windows. 
But it says this:

Hardware sensors are read using external tools. Supported tools include 
SpeedFan, CoreTemp and MBM.


-- 
Cheers, Carlos.

[toc] | [prev] | [next] | [standalone]

#18502

Carlos E.R. wrote:

> I use a tool that is designed for Linux, but also exists, with 
> limitations, for Windows.
> 
> https://en.wikipedia.org/wiki/GKrellM

I've used it in the past, but there's not really any equivalent of
lm-sensors/sensors-detect for Windows ...

[toc] | [prev] | [next] | [standalone]

#18508

On Sat, 4/19/2025 8:36 AM, Carlos E.R. wrote:
> On 2025-04-19 14:04, Frank Slootweg wrote:
>>    IIRC, SpeedFan [1] can show temperatures (and fan speed). Don't know
>> if it can do both at the same time.
> 
> I use a tool that is designed for Linux, but also exists, with limitations, for Windows.
> 
> https://en.wikipedia.org/wiki/GKrellM
> 
> Home page: <https://gkrellm.srcbox.net/>. There is a link for Windows. But it says this:
> 
> Hardware sensors are read using external tools. Supported tools include SpeedFan, CoreTemp and MBM.
> 

There seems to be a pattern emerging. Developers are willing to waste
weeks of their time, developing rich-looking GUIs. But nobody wants
to work on reverse engineering the register level interface on the
latest SuperIO.

   Paul

[toc] | [prev] | [next] | [standalone]

#18468

Andy Burns <usenet@andyburns.uk> wrote:

> Anyway, I've always been a bit suspicious of the MemoryCompression 
> service, so I'm experimenting with it turned off for a bit.
> (powershell disable-mmagent -mc)

As far as memory compression goes, it might has some value on low-end
computers.  See the following about that wonderful OS "feature".

https://www.makeuseof.com/windows-memory-compression-guide/

Even if the data is held entirely in memory, like a buffer or cache,
compression takes CPU cycles, and more CPU cycles generate more heat.
Every read will have to decompress the data before it can be used.
Every write will have to compress the data.  That is overhead on the
CPU.

Apparently the Memory Compression service is always active.  I have 64
GB of RAM, and don't use anywhere near that much at any time.  Yet Task
Manager's Performance tab under Memory shows "Inuse: 6.3 GB (204 MB)
Available: 52.0 GB" which is the amount of RAM in use with the
compressed size in parenthesis, and the last is how much is free.  No
idea why anything in RAM got compressed considering over 50 GB of RAM is
unused.

I was going to ask how much free RAM you have, but I have gobs, and RAM
compression is still getting used.  Because compression is involved
means there are more CPU cycles for the same amount of data, so there
will be more heat with compression enabled.  Considering the overhead
involved to compress on write, and decompress on read, with lots more
CPU cycles to do both, seems another Microsoft gimmick most likely to
get their OS working on low-end computers (those with not much RAM).
You mentioned how much physical RAM you have in your computers, but not
how much is free after loading whatever programs you often use
concurrently.  So, yep, I'll ask: how much free RAM do you have?

I've seen some gamers noting they go from 60 fps to 70 fps with memory
compression disabled.  Some users have reported a whopping change from
30 fps to 120 fps.  3D graphics users have noted the app is slower when
memory compression is enabled.  Less overhead in CPU cycles to compress
on write and decompress on read.  Impact would be affected a lot by the
performance of the CPU: a low-end slow CPU would take longer to compress
and decompress.

https://www.elevenforum.com/t/enable-or-disable-memory-compression-in-windows-10-and-windows-11.3555/

With 64 GB of RAM (often only 10% of it in use), and 1 TB SSD NVMe, 2 TB
SSD, and 4 TB spinner all of which are less than half occupied, or much
less, I certainly don't need RAM and disk compression trying to save
resources that far exceed the demand.  Considering how memory
compression is not saving me from some low-memory disaster, I'll be
disabling it.

get-mmagent

in Powershell (with admin privs) shows it is currently enabled on my
desktop PC.  I see no point in adding all that CPU overhead for
compression when I have gobs of free RAM.

disable-mmagent -mc

Even if I get no performance boost from not compressing into RAM, I'd
rather the CPU run cooler by not wasting cycle on superfluous
compression.

[toc] | [prev] | [next] | [standalone]

#18469

On 2025-04-17 22:07, VanguardLH wrote:

...

> I was going to ask how much free RAM you have, but I have gobs, and RAM
> compression is still getting used.  Because compression is involved
> means there are more CPU cycles for the same amount of data, so there
> will be more heat with compression enabled.  Considering the overhead
> involved to compress on write, and decompress on read, with lots more
> CPU cycles to do both, seems another Microsoft gimmick most likely to
> get their OS working on low-end computers (those with not much RAM).
> You mentioned how much physical RAM you have in your computers, but not
> how much is free after loading whatever programs you often use
> concurrently.  So, yep, I'll ask: how much free RAM do you have?
> 
> I've seen some gamers noting they go from 60 fps to 70 fps with memory
> compression disabled.  Some users have reported a whopping change from
> 30 fps to 120 fps.  3D graphics users have noted the app is slower when
> memory compression is enabled.  Less overhead in CPU cycles to compress
> on write and decompress on read.  Impact would be affected a lot by the
> performance of the CPU: a low-end slow CPU would take longer to compress
> and decompress.

IMHO, memory compression should not be active while there are lots of 
free ram available.

Unless, unused memory can be powered off. Can it? I have not heard of 
such a feature, but it might exist. :-?

-- 
Cheers, Carlos.

[toc] | [prev] | [next] | [standalone]

#18475

"Carlos E.R." <robin_listas@es.invalid> wrote:

> IMHO, memory compression should not be active while there are lots of 
> free ram available.
> 
> Unless, unused memory can be powered off. Can it? I have not heard of 
> such a feature, but it might exist. :-?

Considering how addressing accesses across modules, I doubt it.  Having
to power back up would severely slow memory access speed.  Takes time to
wake up, and these days it's even longer for me.

You can adjust clock rate voltage, CAS latency, and some other timing
tweaks all of which are for overclocking.  I don't overclock outside
what the module specifies itself via XMP Profile.  I just let SPD from
the chip decide the memory settings in BIOS.  I didn't think XMP Profile
was enabled, but after investigating how it works, and how the chip
maker specifies those tweaks within the module which should be stable, I
went into BIOS to enable XMP Profile only to find that is the default,
so it was already enabled.  My mobo only supports one XMP Profile (XMP
2.0 only since I have DDR4, no XMP 3.0 since I don't have DDR5), so my
choice was on or off, not which one to use.  Since I had been running
for several years with it unknowingly enabled, I kept running it that
way.  You can use CPU-Z's SPD tab to check if a memory module is using
XMP, and which version.  Pick a "slot" (memory module), and look at the
"SPD ext." setting.  It's either on or off for all memory.

https://www.crucial.com/articles/about-memory/what-is-xmp
https://www.pcgamer.com/what-are-xmp-profiles-and-how-do-i-use-them/

Other than XMP Profile, I don't don't do overclocking.  I use the
optimal setting (default) specified by the SPD in the memory module.
When I first built the box, I tested with some overclocking to see how
much of a boost I'd get in performance.  In a couple video games, the
frame rate when up a little, like 5 fps, but it was already more than
high enough to show a smooth display during game play.  The extra heat
wasn't worth the tiny boost, and I didn't like upping the voltage.  Heat
is the nemesis of silicon.

I don't recall ever seeing some power management scheme that would power
off or put into low-power mode whatever modules did not have accessed
addresses after some threshold of idle addressing.  How would memory
know it didn't have something written to it by some process that would
expect immediate access, but after some timeout threshold?  However, I
did find:

https://edc.intel.com/content/www/us/en/design/ipla/software-development-platforms/client/platforms/alder-lake-desktop/12th-generation-intel-core-processors-datasheet-volume-1-of-2/009/memory-power-savings-technologies/

RAM doesn't retain its content, so it cannot be powered off; else, you'd
lose everything stored in RAM.  Low-power mode might work, but it would
still have to exceed the minimal voltage needed to operate the RAM to
ensure bits weren't lost.

https://www.intel.com/content/www/us/en/support/articles/000006619/processors/intel-core-processors.html

That mentions the C3 and C4 power states, but for mobile platforms.
Saving battery power is crucial to up-time for mobiles.  The first Intel
article mentions having to keep cycling the memory through refreshes to
keep alive its content during C3 (deep sleep).  Okay, but it takes a
moment to come out of sleep mode, and that would be highly impactful to
the performance and responsiveness of the OS and apps if the computer
were constantly diving into deep sleep to immediately come out of it.

[toc] | [prev] | [next] | [standalone]

#18482

On 2025-04-18 02:30, VanguardLH wrote:
> "Carlos E.R." <robin_listas@es.invalid> wrote:
> 
>> IMHO, memory compression should not be active while there are lots of
>> free ram available.
>>
>> Unless, unused memory can be powered off. Can it? I have not heard of
>> such a feature, but it might exist. :-?
> 
> Considering how addressing accesses across modules, I doubt it.  Having
> to power back up would severely slow memory access speed.  Takes time to
> wake up, and these days it's even longer for me.
> 
> You can adjust clock rate voltage, CAS latency, and some other timing
> tweaks all of which are for overclocking.  I don't overclock outside
> what the module specifies itself via XMP Profile.  I just let SPD from
> the chip decide the memory settings in BIOS.  I didn't think XMP Profile
> was enabled, but after investigating how it works, and how the chip
> maker specifies those tweaks within the module which should be stable, I
> went into BIOS to enable XMP Profile only to find that is the default,
> so it was already enabled.  My mobo only supports one XMP Profile (XMP
> 2.0 only since I have DDR4, no XMP 3.0 since I don't have DDR5), so my
> choice was on or off, not which one to use.  Since I had been running
> for several years with it unknowingly enabled, I kept running it that
> way.  You can use CPU-Z's SPD tab to check if a memory module is using
> XMP, and which version.  Pick a "slot" (memory module), and look at the
> "SPD ext." setting.  It's either on or off for all memory.
> 
> https://www.crucial.com/articles/about-memory/what-is-xmp
> https://www.pcgamer.com/what-are-xmp-profiles-and-how-do-i-use-them/
> 
> Other than XMP Profile, I don't don't do overclocking.  I use the
> optimal setting (default) specified by the SPD in the memory module.
> When I first built the box, I tested with some overclocking to see how
> much of a boost I'd get in performance.  In a couple video games, the
> frame rate when up a little, like 5 fps, but it was already more than
> high enough to show a smooth display during game play.  The extra heat
> wasn't worth the tiny boost, and I didn't like upping the voltage.  Heat
> is the nemesis of silicon.
> 

I never overclock. My board is designed for overclocking, but I just use 
one of the defaults.

> I don't recall ever seeing some power management scheme that would power
> off or put into low-power mode whatever modules did not have accessed
> addresses after some threshold of idle addressing.  How would memory
> know it didn't have something written to it by some process that would
> expect immediate access, but after some timeout threshold?  However, I
> did find:

The operating system would know, and decide not to use part of the 
memory. Switch it off, no contents to worry about losing. But I suspect 
the power saved would be minimal.

> 
> https://edc.intel.com/content/www/us/en/design/ipla/software-development-platforms/client/platforms/alder-lake-desktop/12th-generation-intel-core-processors-datasheet-volume-1-of-2/009/memory-power-savings-technologies/
> 
> RAM doesn't retain its content, so it cannot be powered off; else, you'd
> lose everything stored in RAM.  Low-power mode might work, but it would
> still have to exceed the minimal voltage needed to operate the RAM to
> ensure bits weren't lost.
> 
> https://www.intel.com/content/www/us/en/support/articles/000006619/processors/intel-core-processors.html
> 
> That mentions the C3 and C4 power states, but for mobile platforms.
> Saving battery power is crucial to up-time for mobiles.  The first Intel
> article mentions having to keep cycling the memory through refreshes to
> keep alive its content during C3 (deep sleep).  

Right.

> Okay, but it takes a
> moment to come out of sleep mode, and that would be highly impactful to
> the performance and responsiveness of the OS and apps if the computer
> were constantly diving into deep sleep to immediately come out of it.

Not important, because switching a memory module on would happen when 
loading more software, or when some application demands a lot more 
memory. This would be a rare event.

It might be interesting for phones and other battery operated devices.

You make me wonder: mobile phones use dynamic ram, or static ram?

I think calculators use static.

-- 
Cheers, Carlos.

[toc] | [prev] | [next] | [standalone]

#18489

"Carlos E.R." <robin_listas@es.invalid> wrote:

> The operating system would know, and decide not to use part of the 
> memory. Switch it off, no contents to worry about losing. But I suspect 
> the power saved would be minimal.

I have 64 GB of RAM.  That would be a hell of a lot of writes to
remember (64 GB) to know if some memory had been written in the past.
With my 64 GB of RAM, another 64 GB of RAM or cache on very slow disk
would be need to track which addresses had been written in the RAM.  Or,
alternatively, I could only use 32 GB of my RAM, because the other 32 GB
was used for tracking past writes.  

This history cache to record which addresses were accessed wouldn't need
to be as big as the main RAM that stores the data, or maybe a bit
mapping could be used to equate an address in RAM to a bit in history
RAM, but the history would have to record all writes however long the
computer had been powered up, including sleep modes.  More RAM would be
needed to track writes to main RAM.  You certainly wouldn't use a
history cache on a drive.

>> Okay, but it takes a moment to come out of sleep mode, and that would
>> be highly impactful to the performance and responsiveness of the OS
>> and apps if the computer were constantly diving into deep sleep to
>> immediately come out of it.
> 
> Not important, because switching a memory module on would happen when 
> loading more software, or when some application demands a lot more 
> memory. This would be a rare event.

No, the moment any process needed a single byte from a memory module
would require the memory module to be instantly turned on from sleep.
That includes the OS along with any services and backgrounded programs.
Yes, executables have an image loaded into memory when they get loaded,
but memory is also accessed by running processes.  ALL processes would
have to get suspended to prevent any memory access during sleep.  Well,
that's when sleeping, and neither you nor the computer can do anything,
so it appears all memory modules can go into a partial reduced power
mode, but require constant refreshing and voltage can never be reduced.
The constant refreshing to retain content in memory and required voltage
means the memory module doesn't sleep, it just has a lower duty cycle to
reduce its power consumption, but then DRAM doesn't consume much power.  

> It might be interesting for phones and other battery operated devices.
> 
> You make me wonder: mobile phones use dynamic ram, or static ram?
> 
> I think calculators use static.

Static RAM holds its data as long as it has power.  Does not require
refreshing versus dynamic RAM where data decays in seconds and needs
constant refreshing.  SRAM is faster than DRAM, but not as dense as
DRAM.  SRAM is used in cell phones and wearables, and as a cache for the
CPU registers (i.e., microprocessors have SRAM), but then nowhere near
as much memory is needed as for main system memory where programs are
loaded and data is stored.  

There is also NVSRAM where batteries are impractical to preserve data,
like in medical equipment where data loss is intolerable from a power
loss.  I've never looked into PSRAM (Pseudostatic SRAM) which puts the
refresh logic inside the memory module.

I don't see SRAM sold at Newegg.  I've never had a desktop build nor
pried into a laptop that had sockets for SRAM, just for DRAM.  I did
find SRAM sold online, like at:

https://www.mouser.com/c/semiconductors/memory-ics/sram/

4 GB SRAM (134,217,728 x 32) was the largest they list.  They didn't
show pricing since it was out of stock.  It is a BGA package, so has to
get soldered which reflects its specialty use, not slid into a socket.

I suspect Paul, based on the technical detail of his posts, has dug more
into where SRAM might get used.

[toc] | [prev] | [next] | [standalone]

#18493

On 2025-04-18 18:13, VanguardLH wrote:
> "Carlos E.R." <robin_listas@es.invalid> wrote:
> 
>> The operating system would know, and decide not to use part of the
>> memory. Switch it off, no contents to worry about losing. But I suspect
>> the power saved would be minimal.
> 
> I have 64 GB of RAM.  That would be a hell of a lot of writes to
> remember (64 GB) to know if some memory had been written in the past.
> With my 64 GB of RAM, another 64 GB of RAM or cache on very slow disk
> would be need to track which addresses had been written in the RAM.  Or,
> alternatively, I could only use 32 GB of my RAM, because the other 32 GB
> was used for tracking past writes.
> 
> This history cache to record which addresses were accessed wouldn't need
> to be as big as the main RAM that stores the data, or maybe a bit
> mapping could be used to equate an address in RAM to a bit in history
> RAM, but the history would have to record all writes however long the
> computer had been powered up, including sleep modes.  More RAM would be
> needed to track writes to main RAM.  You certainly wouldn't use a
> history cache on a drive.

No, if the feature existed it would switch on/off large blocks of 
memory, probably entire modules or chips.


> 
>>> Okay, but it takes a moment to come out of sleep mode, and that would
>>> be highly impactful to the performance and responsiveness of the OS
>>> and apps if the computer were constantly diving into deep sleep to
>>> immediately come out of it.
>>
>> Not important, because switching a memory module on would happen when
>> loading more software, or when some application demands a lot more
>> memory. This would be a rare event.
> 
> No, the moment any process needed a single byte from a memory module
> would require the memory module to be instantly turned on from sleep.

No. It would happen only when a program wants to assign, for instance, a 
block of 100 MB of memory, and the operating system decides whether to 
use swap or activate more ram.

Accessing a byte? No, that byte is already assigned and live.

A ram module can not be switched off while any code holds addresses on it.

> That includes the OS along with any services and backgrounded programs.
> Yes, executables have an image loaded into memory when they get loaded,
> but memory is also accessed by running processes.  ALL processes would
> have to get suspended to prevent any memory access during sleep.  Well,
> that's when sleeping, and neither you nor the computer can do anything,
> so it appears all memory modules can go into a partial reduced power
> mode, but require constant refreshing and voltage can never be reduced.
> The constant refreshing to retain content in memory and required voltage
> means the memory module doesn't sleep, it just has a lower duty cycle to
> reduce its power consumption, but then DRAM doesn't consume much power.
> 
>> It might be interesting for phones and other battery operated devices.
>>
>> You make me wonder: mobile phones use dynamic ram, or static ram?
>>
>> I think calculators use static.
> 
> Static RAM holds its data as long as it has power.  Does not require
> refreshing versus dynamic RAM where data decays in seconds and needs
> constant refreshing.  SRAM is faster than DRAM, but not as dense as
> DRAM.  SRAM is used in cell phones and wearables, and as a cache for the
> CPU registers (i.e., microprocessors have SRAM), but then nowhere near
> as much memory is needed as for main system memory where programs are
> loaded and data is stored.
> 
> There is also NVSRAM where batteries are impractical to preserve data,
> like in medical equipment where data loss is intolerable from a power
> loss.  I've never looked into PSRAM (Pseudostatic SRAM) which puts the
> refresh logic inside the memory module.
> 
> I don't see SRAM sold at Newegg.  I've never had a desktop build nor
> pried into a laptop that had sockets for SRAM, just for DRAM.  I did
> find SRAM sold online, like at:
> 
> https://www.mouser.com/c/semiconductors/memory-ics/sram/
> 
> 4 GB SRAM (134,217,728 x 32) was the largest they list.  They didn't
> show pricing since it was out of stock.  It is a BGA package, so has to
> get soldered which reflects its specialty use, not slid into a socket.
> 
> I suspect Paul, based on the technical detail of his posts, has dug more
> into where SRAM might get used.


-- 
Cheers, Carlos.

[toc] | [prev] | [next] | [standalone]

#18471

VanguardLH wrote:

> As far as memory compression goes, it might has some value on low-end
> computers.  See the following about that wonderful OS "feature".
> 
> https://www.makeuseof.com/windows-memory-compression-guide/

Yes, that was the article I read to find how to turn it off.  did you 
spot the true/false error?

> Even if the data is held entirely in memory, like a buffer or cache,
> compression takes CPU cycles, and more CPU cycles generate more heat.
> Every read will have to decompress the data before it can be used.
> Every write will have to compress the data.  That is overhead on the
> CPU.

yes, I didn't/don't know how much CPU the mem compression is using 
behind the scenes, I suspect it hides that from taskmgr?

I don't know what it doing, but given that memory may contain pointer to 
other memory, if you've compressed it, wrst case youmight have to 
intercept reads, unfrig the pointers before you pass the pages back to 
the unsuspecting app ... that could chew some CPU and cause some heat.

> Apparently the Memory Compression service is always active.  I have 64
> GB of RAM, and don't use anywhere near that much at any time.  Yet Task
> Manager's Performance tab under Memory shows "Inuse: 6.3 GB (204 MB)
> Available: 52.0 GB" which is the amount of RAM in use with the
> compressed size in parenthesis, and the last is how much is free.  No
> idea why anything in RAM got compressed considering over 50 GB of RAM is
> unused.
> 
> I was going to ask how much free RAM you have


usually about 6G free out of 16GB physical.  for some reason I had the 
pagefile set to "auto" and it had allocated 16GB, so I had 32GB virtual, 
I've nowchopped that down to 1GB pagefile.

> but I have gobs, and RAM
> compression is still getting used.  Because compression is involved
> means there are more CPU cycles for the same amount of data, so there
> will be more heat with compression enabled.  Considering the overhead
> involved to compress on write, and decompress on read, with lots more
> CPU cycles to do both, seems another Microsoft gimmick most likely to
> get their OS working on low-end computers (those with not much RAM).
> You mentioned how much physical RAM you have in your computers, but not
> how much is free after loading whatever programs you often use
> concurrently.  So, yep, I'll ask: how much free RAM do you have?

right now (with only FF, TB and HWmonitor, taskmgr as foreground apps) 7.4GB

I have a probably 150-200 tabs open, but nowadays they unload in the 
background much better than they used to.

> I've seen some gamers noting they go from 60 fps to 70 fps with memory
> compression disabled.  Some users have reported a whopping change from
> 30 fps to 120 fps.  3D graphics users have noted the app is slower when
> memory compression is enabled.  Less overhead in CPU cycles to compress
> on write and decompress on read.  Impact would be affected a lot by the
> performance of the CPU: a low-end slow CPU would take longer to compress
> and decompress.
> 
> https://www.elevenforum.com/t/enable-or-disable-memory-compression-in-windows-10-and-windows-11.3555/
> 
> With 64 GB of RAM (often only 10% of it in use), and 1 TB SSD NVMe, 2 TB
> SSD, and 4 TB spinner all of which are less than half occupied, or much
> less, I certainly don't need RAM and disk compression trying to save
> resources that far exceed the demand.  Considering how memory
> compression is not saving me from some low-memory disaster, I'll be
> disabling it.

I've got a dell R760 to tinker with at the moment, not huge amounts of 
memory (we had 256GB or 384GB models as VMware hosts several years ago) 
just 128GB but 2x24core xeons, 12Gbps SAS, couple of SSD to boot from 
and couple of SATA for local backups, 10Gb NICs.

> 
> get-mmagent
> 
> in Powershell (with admin privs) shows it is currently enabled on my
> desktop PC.  I see no point in adding all that CPU overhead for
> compression when I have gobs of free RAM.
> 
> disable-mmagent -mc
> 
> Even if I get no performance boost from not compressing into RAM, I'd
> rather the CPU run cooler by not wasting cycle on superfluous
> compression.


yes, with compression off,this machine is less fan-happy, if I'm 
watching youtube,sure the fan is on, but if firefox is in th ebackground 
and I'm just reading/typing in TB, then the fan shuts off pretty easily, 
whereas before it was constant.

[toc] | [prev] | [next] | [standalone]

#18476

Andy Burns <usenet@andyburns.uk> wrote:

> VanguardLH wrote:
> 
>> As far as memory compression goes, it might has some value on low-end
>> computers.  See the following about that wonderful OS "feature".
>> 
>> https://www.makeuseof.com/windows-memory-compression-guide/
> 
> Yes, that was the article I read to find how to turn it off.  did you 
> spot the true/false error?

No.  I just ran get-mmagent to see the current setting, and expected
True meant on and False meant off.  When I used "disable-mmagent -mc",
the setting changed from True to False.  However, perhaps a reboot is
needed to actually effect the change to the Memory Manager.  Not a
critical feature for me, so eventually I'll get around rebooting the
desktop for some other reason which would then effect this change.

> I have a probably 150-200 tabs open, but nowadays they unload in the 
> background much better than they used to.

Supposedly having inactive tabs reduces their memory footprint to help
keep reduce Firefox's memory consumption.  MS Edge has sleeping tabs,
too, so perhaps all Chromium variants have the feature.

https://support.mozilla.org/en-US/kb/unload-inactive-tabs-save-system-memory-firefox
https://firefox-source-docs.mozilla.org/browser/tabunloader/ (*)

https://www.microsoft.com/en-us/edge/learning-center/how-to-enable-disable-sleeping-tabs-in-microsoft-edge

Apparently some folks don't like the hesitation on awakening a sleeping
tab, especially on mobile platforms, and more so for users having LOTS
of tabs open at once, so they disable the feature.  

For Firefox, you have to dive into the advanced editor (about:config) to
change browser.tabs.unloadOnLowMemory from the default of True to False.
You go to about:unloads to see which tabs are inactive (sleeping).  I'm
not the type to open dozens, or hundreds, or tab to leave them loaded,
plus I don't leave the web browser running when I'm not using it, and I
have 64 GB of RAM, so highly unlikely I would hit a threshold for the
tabs.unloadOnLowMemory trigger.  Plus, I don't load dozens or hundreds
of tabs to bounce between them at a high rate.  I load a tab, I read a
tab, and likely if done then I close the tab.

For Edge (and perhaps other Chromium variants, no having to dig into
some advanced editor.  Under System and Performance in Settings, toggle
the "Save resources with sleeping tabs" option.

I'm sure mobile users would be more interested in sleeping tabs where
both power and memory consumption get reduced for inactive tabs.  On my
Android phone, Firefox has no about:config advanced editor (Mozilla took
it away as they consider their users to be boobs that should not have
access to a gun to shoot themselves in their collective feet); however,
I don't load tons of tabs, so sleeping tabs is something I give a gnat's
fart about.  In fact, I use Firefox Android's Quit menu to actually
unload the web browser rather than leave it running in the background,
plus I configured it, like Firefox Desktop, to purge all its locally
cached data on its exit, so I need to exit Firefox for that option to be
effected.

[toc] | [prev] | [next] | [standalone]

#18479

On Thu, 4/17/2025 5:12 PM, Andy Burns wrote:
> VanguardLH wrote:
> 
>> As far as memory compression goes, it might has some value on low-end
>> computers.  See the following about that wonderful OS "feature".
>>
>> https://www.makeuseof.com/windows-memory-compression-guide/
> 
> Yes, that was the article I read to find how to turn it off.  did you spot the true/false error?
> 
>> Even if the data is held entirely in memory, like a buffer or cache,
>> compression takes CPU cycles, and more CPU cycles generate more heat.
>> Every read will have to decompress the data before it can be used.
>> Every write will have to compress the data.  That is overhead on the
>> CPU.
> 
> yes, I didn't/don't know how much CPU the mem compression is using behind the scenes, I suspect it hides that from taskmgr?
> 
> I don't know what it doing, but given that memory may contain pointer to other memory, if you've compressed it, wrst case youmight have to intercept reads, unfrig the pointers before you pass the pages back to the unsuspecting app ... that could chew some CPU and cause some heat.
> 
>> Apparently the Memory Compression service is always active.  I have 64
>> GB of RAM, and don't use anywhere near that much at any time.  Yet Task
>> Manager's Performance tab under Memory shows "Inuse: 6.3 GB (204 MB)
>> Available: 52.0 GB" which is the amount of RAM in use with the
>> compressed size in parenthesis, and the last is how much is free.  No
>> idea why anything in RAM got compressed considering over 50 GB of RAM is
>> unused.
>>
>> I was going to ask how much free RAM you have
> 
> 
> usually about 6G free out of 16GB physical.  for some reason I had the pagefile set to "auto" and it had allocated 16GB, so I had 32GB virtual, I've nowchopped that down to 1GB pagefile.
> 
>> but I have gobs, and RAM
>> compression is still getting used.  Because compression is involved
>> means there are more CPU cycles for the same amount of data, so there
>> will be more heat with compression enabled.  Considering the overhead
>> involved to compress on write, and decompress on read, with lots more
>> CPU cycles to do both, seems another Microsoft gimmick most likely to
>> get their OS working on low-end computers (those with not much RAM).
>> You mentioned how much physical RAM you have in your computers, but not
>> how much is free after loading whatever programs you often use
>> concurrently.  So, yep, I'll ask: how much free RAM do you have?
> 
> right now (with only FF, TB and HWmonitor, taskmgr as foreground apps) 7.4GB
> 
> I have a probably 150-200 tabs open, but nowadays they unload in the background much better than they used to.
> 
>> I've seen some gamers noting they go from 60 fps to 70 fps with memory
>> compression disabled.  Some users have reported a whopping change from
>> 30 fps to 120 fps.  3D graphics users have noted the app is slower when
>> memory compression is enabled.  Less overhead in CPU cycles to compress
>> on write and decompress on read.  Impact would be affected a lot by the
>> performance of the CPU: a low-end slow CPU would take longer to compress
>> and decompress.
>>
>> https://www.elevenforum.com/t/enable-or-disable-memory-compression-in-windows-10-and-windows-11.3555/
>>
>> With 64 GB of RAM (often only 10% of it in use), and 1 TB SSD NVMe, 2 TB
>> SSD, and 4 TB spinner all of which are less than half occupied, or much
>> less, I certainly don't need RAM and disk compression trying to save
>> resources that far exceed the demand.  Considering how memory
>> compression is not saving me from some low-memory disaster, I'll be
>> disabling it.
> 
> I've got a dell R760 to tinker with at the moment, not huge amounts of memory (we had 256GB or 384GB models as VMware hosts several years ago) just 128GB but 2x24core xeons, 12Gbps SAS, couple of SSD to boot from and couple of SATA for local backups, 10Gb NICs.
> 
>>
>> get-mmagent
>>
>> in Powershell (with admin privs) shows it is currently enabled on my
>> desktop PC.  I see no point in adding all that CPU overhead for
>> compression when I have gobs of free RAM.
>>
>> disable-mmagent -mc
>>
>> Even if I get no performance boost from not compressing into RAM, I'd
>> rather the CPU run cooler by not wasting cycle on superfluous
>> compression.
> 
> 
> yes, with compression off,this machine is less fan-happy, if I'm watching youtube,sure the fan is on, but if firefox is in th ebackground and I'm just reading/typing in TB, then the fan shuts off pretty easily, whereas before it was constant.
> 

Process Explorer is the tool you want.

It lists the Memory Compressor.

And if the Memory Compressor is active, *something* is responsible
for the symptoms. A single process can cause this to happen. Your
job is to figure out the root cause.

Memory Compressor is not bugged. Something is telling it to act,
and it is processing its particular flavour of input events.

But part of the problem, is the Task Manager is not a very
honest accounting of machine activity, and the extra two digits
of precision on Process Explorer may help you identify
active artifice (poorly written software, like a Classic Outlook
wrapped in a Webview2 wrapper and "spiking a CPU core").

   Paul

[toc] | [prev] | [next] | [standalone]

#18481

Paul wrote:

> Process Explorer is the tool you want.
> It lists the Memory Compressor.

Yes, I use the entire suite.

until I disabled it (and rebooted) taskmgr listed memory compressor on 
mine, wouldn't show much detail, but then procexp doesn't show as much 
as for a normal process ... yes zillions of clock cycles

> And if the Memory Compressor is active,something is responsible
> for the symptoms. A single process can cause this to happen. Your
> job is to figure out the root cause.

well, i think this slim laptop has marginal cooling, any video will trip 
the fans on, just a couple of programs (FF/TB/LO) along with the 
compressor would be enough to ensure they rarely went off.

> Memory Compressor is not bugged. Something is telling it to act,
> and it is processing its particular flavour of input events.

As far as I'm concerned, it's gone for good, seeing no harm from not 
running it, still have 1/3 of my 16GB available, very likely that 
non-required compression was pushing the CPU over the brink, rather than 
memory usage itself.

> But part of the problem, is the Task Manager is not a very
> honest accounting of machine activity

it's a shame it used to be, but it quite often disagrees with procexp by 
a long way.

  and the extra two digits
> of precision on Process Explorer may help you identify
> active artifice (poorly written software, like a Classic Outlook
> wrapped in a Webview2 wrapper and "spiking a CPU core").
I think it's just the machine and cutting the cpu from compression helps 
it to cope ... my ears appreciate it.

[toc] | [prev] | [next] | [standalone]

#18491

Andy Burns <usenet@andyburns.uk> wrote:
[...]
> well, i think this slim laptop has marginal cooling, any video will trip 
> the fans on, just a couple of programs (FF/TB/LO) along with the 
> compressor would be enough to ensure they rarely went off.

  I understand that your fan problem is mostly gone after disabling
memory compression, so this is just a heads up on your comment:

  My 'slim laptop' (HP Pavilion 15-eh2560nd) shows similar behaviour.
The fan is hardly ever on (meaning audible), but sometimes, for no
apparent reason, it's 'noisy' and when I check at those times, the GPU
temperature (Task Manager only shows the temperature of the integrated
GPU, not of the CPU) is slightly elevated, mid 40s degrees C. Most of
the time I don't even know why the GPU had been in use, but apparently
it had been. When I look, the GPU percentage is always low, single
digit.

  Bottom line: I think what you are/were experiencing is not uncommon
for a modern/current 'slim laptop'.

[...]

[toc] | [prev] | [next] | [standalone]

#18492

Frank Slootweg wrote:

>    Bottom line: I think what you are/were experiencing is not uncommon
> for a modern/current 'slim laptop'.

I can't remember if I had a choice of an i5 model, but I went for the 
i7-11370H probably overkill for the home machine ... I also think the 
firmware upgrades have altered the fan on/slow/off points ...

[toc] | [prev] | [next] | [standalone]

#18470

On 2025-04-17 16:16, Andy Burns wrote:

...

> Memory chips don't know if part of them is allocated to a process or 
> not, right?  Other than maybe some garbage collection, nothing goes 
> around flipping bits for fun, right?
> 
> So how does being allocated, even when sitting idle, influence cooling 
> required?

Memory allocation should have negligible or no effect at all on heat 
generated.

Memory heat depends, AFAIK, solely on memory activity. Reading or 
writing cells.

Unless there is a feature that allows free memory to be powered down or 
not refreshed. I have not heard of it. Considering that a laptop can 
sleep, while it keeps all ram holding its data on battery for some days, 
I don't think that unused ram uses any amount of significant power.

-- 
Cheers, Carlos.

[toc] | [prev] | [next] | [standalone]

#18472

Carlos E.R. wrote:

> Memory heat depends, AFAIK, solely on memory activity. Reading or 
> writing cells.

yes that's what I thought, probably with writes being hungrier than 
reads,but what about refresh, isn't that a constant background 
read/write s it doesn't forget?

> Unless there is a feature that allows free memory to be powered down or 
> not refreshed. I have not heard of it. 

No not heard of it, hypervisors do some clever tricks t spot multiple 
VMs with identical pages, anfd point them all to one copy, but thats to 
free up extra memory for those that need it, not to switch any off

> Considering that a laptop can 
> sleep, while it keeps all ram holding its data on battery for some days, 
> I don't think that unused ram uses any amount of significant power.

[toc] | [prev] | [next] | [standalone]

#18474

On 2025-04-17 23:17, Andy Burns wrote:
> Carlos E.R. wrote:
> 
>> Memory heat depends, AFAIK, solely on memory activity. Reading or 
>> writing cells.
> 
> yes that's what I thought, probably with writes being hungrier than 
> reads,but what about refresh, isn't that a constant background read/ 
> write s it doesn't forget?

Yes, AFAIK refresh is constant.


>> Unless there is a feature that allows free memory to be powered down 
>> or not refreshed. I have not heard of it. 
> 
> No not heard of it, hypervisors do some clever tricks t spot multiple 
> VMs with identical pages, anfd point them all to one copy, but thats to 
> free up extra memory for those that need it, not to switch any off
> 
>> Considering that a laptop can sleep, while it keeps all ram holding 
>> its data on battery for some days, I don't think that unused ram uses 
>> any amount of significant power.
> 
> 


-- 
Cheers, Carlos.

[toc] | [prev] | [next] | [standalone]

#18480

On Thu, 4/17/2025 5:17 PM, Andy Burns wrote:
> Carlos E.R. wrote:
> 
>> Memory heat depends, AFAIK, solely on memory activity. Reading or writing cells.
> 
> yes that's what I thought, probably with writes being hungrier than reads,but what about refresh, isn't that a constant background read/write s it doesn't forget?
> 
>> Unless there is a feature that allows free memory to be powered down or not refreshed. I have not heard of it. 
> 
> No not heard of it, hypervisors do some clever tricks t spot multiple VMs with identical pages, anfd point them all to one copy, but thats to free up extra memory for those that need it, not to switch any off
> 
>> Considering that a laptop can sleep, while it keeps all ram holding its data on battery for some days, I don't think that unused ram uses any amount of significant power.
> 
> 

When you get a memory power spec, the "representative power" is via
the "Industry Standard Cycle Mix".

Each cycle type has a power consumption.

A "No Operation" on the command bus, still burns power, but that is
"toggle power". The clock signal has changed state twice, and internal
logic that changes state, dissipates a tiny amount of heat.

The next cycle type up from that, is "Refresh". Refresh uses a little
more power. AutoRefresh (computer sleeping), uses 1 watt over 16 chips.
the command bus during AutoRefresh is silent (the memory controller is
switched off during sleep). All of the power there, is counters inside
the module, incrementing Row and Column address, as memory locations
are visited every 64 milliseconds and recharged.

"Read" and "Write" cycle types, occur in bursts. The CPU likes to do
cache line sized accesses (64 bytes at a time). Two DIMMs in dual
channel are 16 bytes wide. A burst-of-4 is done in that case, to
complete a cache line fill or a cache line removal. If you run
a single DIMM, the memory controller uses a burst-of-8 for the cache
line activity. These cycles are very expensive, compared to the previous
cycles. But if you only do 4 of those, per 100 cycle interval,
or 8 of those per 100 cycle interval, the RAM only gets pleasantly warm.

The memory supports worse modes than that, but if you as an engineer
decide to use them, if you hammer the RAM, then be prepared to add
cooling as appropriate. I think you can do something like 256 cycles
back-to-back as a burst, which might be used for an expensive RAM drive
with a PCIe interface.

There are a couple transaction types. A page can be closed or a page
can be open. The cycle time for an open page, is fewer cycles than
if the page has to be opened. Maybe one is 60 cycles, the other 100 cycles.
In usage patterns like Memtest, you would expect a few of the tests
to be using a lot of open pages. The RAM runs a bit warmer, if you
check a Memtest run and use your finger to check temps.

In a desktop, if you use "fat" modules with heat spreaders, counter
intuitively, that's bad. There is no air channel between DIMMs when
you do that. The two modules in the middle of the four module cluster,
can be warmer than the outside ones.

But unlike GDDR7 on a video card, running at 95C or 100C, or L3 cache
in an AMD processor sandwiched underneath a CCX, the DRAM modules we
use for the most part, have a cool and uneventful life.

It is when you over-volt modules, they get hot. Some people might
be running Winbond modules, with a nominal 2.5V operating voltage,
ad 3.6V. Such kits came with blowers to fit over the RAM modules.
Part of that heat exists, because the DRAM chips have voltage
regulators inside (generating internal voltages for certain parts
of the design), and linear regulators are not efficient by any means,
and when really excessive amounts of voltage are used, those modules
are going to get *hot*. Burn yourself hot.

Ever since Intel started listing relative limited overvolt ranges,
the practice of incinerating RAM has kinda stopped. A 1.5V RAM
standard, had 1.65V as the "Absolute Max". After a few failures,
some of the users started recognizing those datasheet items as important.
And that amount of overvolt, is unlikely to even need those fancy
dragon-shaped metal covers with sharp points for your fingers.
Probably all the modules could be sold with just the memory chips
and their surface area for cooling. And the thinner modules would
have a tiny air channel between them.

On multiple occasions, Intel advertised the possibility of RAM overheat
and talked of "cycle rate limiters", as a means to avoid sticky situations.
The industry response, was you could hear crickets. Absolutely nobody
panicked. Today, you might find that memory modules do have
temperature sensors (DDR5 maybe), but there might not be any consumer
software to read that out.

The CPU power, there is a TDP figure. You can use half the TDP,
by running one core railed. The reason for that, is the racetrack
ring with the core voltage on it, is raised for all cores, and the
idle cores still leak a bit of power. When additional cores engage,
the incremental power penalty isn't all that great.

But you can see from that observation, that the memory compressor
running on one core, that's not going to be good for battery life.
Good battery life, comes from bursty CPU activity, where the CPU
voltage drops in microseconds, back to the power saving voltage,
and your battery lasts for hours and hours.

And now I have to go over to the machine on the bench, and figure
out what the hell it is doing in the background :-/ I have the USB
stick with the tools on it, to load up the machine and start looking.
This is a Windows 10, with barely any software installed on it. There is
no WSL2 on it, and I don't think there is a VBox or a VMWare on it.

   Paul

[toc] | [prev] | [standalone]

Page 2 of 2 — ← Prev page 1 [2]

Back to top | Article view | alt.comp.os.windows-11

csiph-web