From Newsgroup: comp.os.linux.advocacy

My new (pre-owned) DDR3 RAM arrived today. Now my 12-y.o.
system is operating with all slots full and 32 Gb of memory.

Here is the relevant portion of the kernel boot log:

Jun 30 12:10:15 (none) kernel: Kernel compiled without mitigations, ignoring 'mitigations'; system may still be vulnerable
Jun 30 12:10:15 (none) kernel: NX (Execute Disable) protection: active
Jun 30 12:10:15 (none) kernel: APIC: Static calls initialized
Jun 30 12:10:15 (none) kernel: DMI: SMBIOS 2.7 present.
Jun 30 12:10:15 (none) kernel: DMI: Gigabyte Technology Co., Ltd. H97-D3H/H97-D3H-CF, BIOS F5 06/26/2014
Jun 30 12:10:15 (none) kernel: DMI: Memory slots populated: 4/4
Jun 30 12:10:15 (none) kernel: tsc: Fast TSC calibration using PIT
Jun 30 12:10:15 (none) kernel: tsc: Detected 3399.837 MHz processor
Jun 30 12:10:15 (none) kernel: e820: update [mem 0x00000000-0x00000fff] System RAM ==> device reserved
Jun 30 12:10:15 (none) kernel: e820: remove [mem 0x000a0000-0x000fffff] System RAM
Jun 30 12:10:15 (none) kernel: last_pfn = 0x81e000 max_arch_pfn = 0x400000000 Jun 30 12:10:15 (none) kernel: total RAM covered: 32736M

It's time to pop the champagne and celebrate as GNU/Linux and
only GNU/Linux breathes fantastic life into an "ancient"
machine.

This 12-y.o. and highly OPTIMIZED system will rival anything
more "modern."

--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.advocacy

Distro Lackey wrote:

My new (pre-owned) DDR3 RAM arrived today. Now my 12-y.o.
system is operating with all slots full and 32 Gb of memory.

Here is the relevant portion of the kernel boot log:

Jun 30 12:10:15 (none) kernel: Kernel compiled without mitigations, ignoring 'mitigations'; system may still be vulnerable
Jun 30 12:10:15 (none) kernel: NX (Execute Disable) protection: active
Jun 30 12:10:15 (none) kernel: APIC: Static calls initialized
Jun 30 12:10:15 (none) kernel: DMI: SMBIOS 2.7 present.
Jun 30 12:10:15 (none) kernel: DMI: Gigabyte Technology Co., Ltd. H97-D3H/H97-D3H-CF, BIOS F5 06/26/2014
Jun 30 12:10:15 (none) kernel: DMI: Memory slots populated: 4/4
Jun 30 12:10:15 (none) kernel: tsc: Fast TSC calibration using PIT
Jun 30 12:10:15 (none) kernel: tsc: Detected 3399.837 MHz processor
Jun 30 12:10:15 (none) kernel: e820: update [mem 0x00000000-0x00000fff] System RAM ==> device reserved
Jun 30 12:10:15 (none) kernel: e820: remove [mem 0x000a0000-0x000fffff] System RAM
Jun 30 12:10:15 (none) kernel: last_pfn = 0x81e000 max_arch_pfn = 0x400000000 >Jun 30 12:10:15 (none) kernel: total RAM covered: 32736M

It's time to pop the champagne and celebrate as GNU/Linux and
only GNU/Linux breathes fantastic life into an "ancient"
machine.

This 12-y.o. and highly OPTIMIZED system will rival anything
more "modern."

Good job, Fabian!
--
"MS is going to take over linux. Soon we'll have a real linux
desktop." - some piece of shit that called itself "GreyCloud"
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.advocacy

On Tue, 30 Jun 2026 13:00:54 -0500, chrisv wrote:

Good job, Fabian!

It wasn't much of a job. Just plug 'em in and then boot
it up.

Of course, one must select a properly matched RAM set
and that's where the utility "dmidecode" comes into
play:

<https://www.nongnu.org/dmidecode/>

Dmidecode reads the BIOS which has read the actual memory
specs from the installed hardware.

Here is the relevant dmidecode output for this machine:

Memory Device
Array Handle: 0x0042
Error Information Handle: Not Provided
Total Width: 64 bits
Data Width: 64 bits
Size: 8 GiB
Form Factor: DIMM
Set: None
Locator: ChannelA-DIMM0
Bank Locator: BANK 0
Type: DDR3
Type Detail: Synchronous
Speed: 1333 MT/s
Manufacturer: 1315
Serial Number: A012B021
Asset Tag: 9876543210
Part Number: BLT8G3D1608DT1TX0.
Rank: 2
Configured Memory Speed: 1333 MT/s


I also have kept the original digital manuals that
came with this motherboard and these also indicate
a recommended memory of Crucial BLT8G3D1608DT1TX0.

Next came udm14.com (Google AI free) that led me to ebay
and the pre-owned sticks.

I have a powerful fan that blows tons of cooling air
across this naked motherboard. My CPU, memory, and hard
drives NEVER will exceed 37C and it is HEAT that destroys
components.

--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.advocacy

At Tue, 30 Jun 2026 16:27:07 +0000, Distro Lackey <dl@lackey.com> wrote:

My new (pre-owned) DDR3 RAM arrived today. Now my 12-y.o.
system is operating with all slots full and 32 Gb of memory.

Here is the relevant portion of the kernel boot log:

Jun 30 12:10:15 (none) kernel: Kernel compiled without mitigations, ignoring 'mitigations'; system may still be vulnerable
Jun 30 12:10:15 (none) kernel: NX (Execute Disable) protection: active
Jun 30 12:10:15 (none) kernel: APIC: Static calls initialized
Jun 30 12:10:15 (none) kernel: DMI: SMBIOS 2.7 present.
Jun 30 12:10:15 (none) kernel: DMI: Gigabyte Technology Co., Ltd. H97-D3H/H97-D3H-CF, BIOS F5 06/26/2014
Jun 30 12:10:15 (none) kernel: DMI: Memory slots populated: 4/4
Jun 30 12:10:15 (none) kernel: tsc: Fast TSC calibration using PIT
Jun 30 12:10:15 (none) kernel: tsc: Detected 3399.837 MHz processor
Jun 30 12:10:15 (none) kernel: e820: update [mem 0x00000000-0x00000fff] System RAM ==> device reserved
Jun 30 12:10:15 (none) kernel: e820: remove [mem 0x000a0000-0x000fffff] System RAM
Jun 30 12:10:15 (none) kernel: last_pfn = 0x81e000 max_arch_pfn = 0x400000000 Jun 30 12:10:15 (none) kernel: total RAM covered: 32736M

It's time to pop the champagne and celebrate as GNU/Linux and
only GNU/Linux breathes fantastic life into an "ancient"
machine.

This 12-y.o. and highly OPTIMIZED system will rival anything
more "modern."

So what is 32GB/$(nproc)?

You need to take that into account for each gcc process
when you run make -j $(nproc).

Here I have this:

$ nproc
64

So about 4GB/processor. Good for gcc.

(The new machine has 512GB/128, so still 4GB/processor.)
--
-v System76 Thelio Mega v1.1 x86_64 Mem: 258G
OS: Linux 7.1.2 D: Mint 22.3 DE: Xfce 4.18 (X11)
NVIDIA GeForce RTX 3090Ti (24G) (610.43.02)
"Do witches run spell checkers?"
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.advocacy

Le 30-06-2026, Distro Lackey <dl@lackey.com> a écrit :

It's time to pop the champagne and celebrate as GNU/Linux and
only GNU/Linux breathes fantastic life into an "ancient"
machine.

Well, it's not that new anymore because you replaced some old part by
new ones.

This 12-y.o. and highly OPTIMIZED system will rival anything
more "modern."

In your dreams.

First you proved more than once that you can optimize your Linux.
Second, as long as you'll keep old slow HDD drives, you will never be
able to compete again a system with SSD.

As long as you stare as you black useless screen with nothing on it, you
can believe your computer is fast. But the only way for the garbage, you
call a highly optimized system, to compete with a modern computer is when
that competing computer is shut down.
--
Si vous avez du temps à perdre :
https://scarpet42.gitlab.io
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.advocacy

On 03 Jul 2026 17:09:54 GMT, Stéphane CARPENTIER wrote:

First you proved more than once that you can optimize your Linux.
Second, as long as you'll keep old slow HDD drives, you will never be
able to compete again a system with SSD.

TALK is cheap. It is ACTION that matters.

Download the scimark4 benchmark:

<https://math.nist.gov/scimark2/scimark4c.zip>

Build and then execute. Report the results (it only
requires 1-2 minutes).

My 12 y.o. OPTIMIZED machine will very likely beat
your modern hardware.

Note also that computation is NOT performed on disj
drives so HDD vs. SSD is irrelevant.


--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.advocacy

Le 03-07-2026, Distro Lackey <dl@lackey.com> a écrit :

On 03 Jul 2026 17:09:54 GMT, Stéphane CARPENTIER wrote:

First you proved more than once that you can optimize your Linux.
Second, as long as you'll keep old slow HDD drives, you will never be
able to compete again a system with SSD.

TALK is cheap. It is ACTION that matters.

Yes, that's why you talk a lot and do nothing.

Download the scimark4 benchmark:

<https://math.nist.gov/scimark2/scimark4c.zip>

Build and then execute. Report the results (it only
requires 1-2 minutes).

My 12 y.o. OPTIMIZED machine will very likely beat
your modern hardware.

I don't have a modern hardware. I've an almost six years old computer.
And it wasn't a very fast one at the time, so I never pretended my
computer would beat a modern computer. I'm pretending your computer
isn't gona beat a modern computer. So the results on my computer are
irrelevant against your claim.

Note also that computation is NOT performed on disj
drives so HDD vs. SSD is irrelevant.

OK, so your test is a useless one. If you can't use a test which looks
like a real case in the real life, it's useless. It will explain why a
computer can look great when it sucks in real usage.

I don't have 500Go of RAM, so my hard drive is useful and it must be as
fast as possible. I have 16Go of RAM, which is well enough for using my computer but isn't enough to avoid using my hard drive.
--
Si vous avez du temps à perdre :
https://scarpet42.gitlab.io
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.advocacy

On 03 Jul 2026 18:54:16 GMT, Stéphane CARPENTIER wrote:

Note also that computation is NOT performed on disk
drives so HDD vs. SSD is irrelevant.

OK, so your test is a useless one. If you can't use a test which looks
like a real case in the real life, it's useless. It will explain why a computer can look great when it sucks in real usage.

I don't have 500Go of RAM, so my hard drive is useful and it must be as
fast as possible. I have 16Go of RAM, which is well enough for using my computer but isn't enough to avoid using my hard drive.

Wrong again, as usual.

Computation is performed in RAM and not on disk drives.
Some software may perform disk I/O during execution but
any disk I/O is usually buffered in RAM.

In conclusion, SDDs are only beneficial for software loading
and start up. SDDs make no great difference unless poorly-written
and bloated software makes heavy use of I/O during execution
that RAM buffering cannot handle.

HDDs are superior to SSDs because they last far longer and don't
require constant "trimming."

Only a total idiot, like YOU, would prefer SSDs.

--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.advocacy

Le 03-07-2026, Distro Lackey <dl@lackey.com> a écrit :

On 03 Jul 2026 18:54:16 GMT, Stéphane CARPENTIER wrote:

Note also that computation is NOT performed on disk
drives so HDD vs. SSD is irrelevant.

OK, so your test is a useless one. If you can't use a test which looks
like a real case in the real life, it's useless. It will explain why a
computer can look great when it sucks in real usage.

I don't have 500Go of RAM, so my hard drive is useful and it must be as
fast as possible. I have 16Go of RAM, which is well enough for using my
computer but isn't enough to avoid using my hard drive.

Wrong again, as usual.

Computation is performed in RAM and not on disk drives.
Some software may perform disk I/O during execution but
any disk I/O is usually buffered in RAM.

In conclusion, SDDs are only beneficial for software loading
and start up. SDDs make no great difference unless poorly-written
and bloated software makes heavy use of I/O during execution
that RAM buffering cannot handle.

That's why you have never been, are not and will never be a scientist. A scientist starts with an explanation or a theory. But, if the facts
contradict that theory/explanation, that scientist looks for a better theory/explanation.

And anyone who has ever used a computer with a SSD will know that it's
faster to use than a computer with a HDD. You forget about the files
used by your applications. For example, when you record some videos
about your failures to use correctly Linux. Guess what: they aren't
stored in RAM.

HDDs are superior to SSDs because they last far longer and don't
require constant "trimming."

In the theory, the SSD have a limited lifespan due to the limited number
of write by sector. In reality, the number of write allowed by sector
expand far beyond a normal use of the SSD. When I have saw too many HDD failures to recognize them as safe.

Only a total idiot, like YOU, would prefer SSDs.

Anyone who used a SSD once in his life can't get back to HDD. It's like
the floppy disk, nobody in his right mind would agree to use them if it
can be avoided. And I'm sure you would use floppy disk if you could,
because you would believe it's cool and makes you looking smart, when
you would just proved how retard you are.
--
Si vous avez du temps à perdre :
https://scarpet42.gitlab.io
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.os.linux.advocacy

On 03 Jul 2026 19:47:42 GMT, Stéphane CARPENTIER wrote:

And anyone who has ever used a computer with a SSD will know that it's
faster to use than a computer with a HDD. You forget about the files
used by your applications. For example, when you record some videos
about your failures to use correctly Linux. Guess what: they aren't
stored in RAM.

Only a deluded idiot would believe so.

As I said, SSDs only allow faster program loading. For example,
whenever I invoke a HUGE program like LibreOffice or Firefox
there can be a lag of several seconds as the bytes are read from
an HDD. With an SSD the program loading would be much faster.

But I don't care about a few seconds of loading time. I care
about reliability and longevity and only an HDD can deliver.
SSDs are a lot more expensive and don't last as long and the
data is always at risk of corruption.

Also, as I already indicated, file I/O in GNU/Linux as always
buffered. Files are stored in RAM until they are explicitly
synced to the disk drive. Furthermore, such buffering (virtual
memory) can be configured by the user.

If I am concerned about video I/O I can even setup a RAM disk
where the file always resides in RAM.

There is absolutely no need for SSDs.

In the theory, the SSD have a limited lifespan due to the limited number
of write by sector. In reality, the number of write allowed by sector
expand far beyond a normal use of the SSD. When I have saw too many HDD failures to recognize them as safe.

SSD hardware can be very good, but superior hardware is also prohibitively expensive ($$$$).

Consumer grade SSDs must use inferior hardware to reduce the price
to a level that is reasonable for the consumer market. Such inferior
hardware is NOT reliable.

Consumer grade SSDs contain a very small amount of superior hardware
to record the location of the failed inferior bits. Unfortunately, the
entire SSD cannot be made from that superior hardware because the cost
would be very, very high.

Consume grade SSDs are a joke. Maybe in the future, when the price of
superior hardware drops, I will acquire an SSD. But now they are a
stupid joke.

--- Synchronet 3.22a-Linux NewsLink 1.2