On 2023-11-13 15:02, Paul wrote:
> On 11/13/2023 7:32 AM, Andy Burns wrote:
>> Frank Slootweg wrote:
>>
>>> Why Windows 10 on a new computer (instead of Windows 11)?
>>
>> That was my initial thought (but I decided not to voice it for no particular reason).
>
> Simple. Two OSes for the price of one. You can upgrade
> any time from W10 to W11.

I would not buy a new computer with Windows 10.

Why?

Well, upgrading would put the effort on me. If there is any problem, it
is me who has to fight and solve it. If it comes with W11, it means the
machine has been tested with W11 and it will work.

If I see a computer sold with W10, I get suspicious. It may be old
hardware, packaged and installed one or two years ago, sold now. If not
packaged, designed time ago, they imaged the install and put it again
and again.

Whether one likes W10 more than 11 is irrelevant.

--
Cheers,
Carlos E.R.

On 11/14/2023 6:54 AM, Carlos E. R. wrote:
> On 2023-11-13 15:02, Paul wrote:
>> On 11/13/2023 7:32 AM, Andy Burns wrote:
>>> Frank Slootweg wrote:
>>>
>>>> Why Windows 10 on a new computer (instead of Windows 11)?
>>>
>>> That was my initial thought (but I decided not to voice it for no particular reason).
>>
>> Simple. Two OSes for the price of one. You can upgrade
>> any time from W10 to W11.
>
> I would not buy a new computer with Windows 10.
>
> Why?
>
> Well, upgrading would put the effort on me. If there is any problem, it is me who has to fight and solve it. If it comes with W11, it means the machine has been tested with W11 and it will work.
>
> If I see a computer sold with W10, I get suspicious. It may be old hardware, packaged and installed one or two years ago, sold now. If not packaged, designed time ago, they imaged the install and put it again and again.
>
> Whether one likes W10 more than 11 is irrelevant.
>

Some people buy equipment, based on finding it has W10.

That makes it a marketing consideration.

Same as when some of the companies ship "new" computers,
where the CPU SKU is three releases old. The parts are
just as worthwhile, but that may allow the retail price
of the box to be lower.

The computer will handle the upgrade from W10 to W11 for you.
It'll even attempt the upgrade, when the upgrade won't work,
then roll the machine back. It's not like you have to
raise a finger to get this to work. No making ISOs required.

Paul

On 2023-11-14 14:43, Paul wrote:
> On 11/14/2023 6:54 AM, Carlos E. R. wrote:
>> On 2023-11-13 15:02, Paul wrote:
>>> On 11/13/2023 7:32 AM, Andy Burns wrote:
>>>> Frank Slootweg wrote:
>>>>
>>>>> Why Windows 10 on a new computer (instead of Windows 11)?
>>>>
>>>> That was my initial thought (but I decided not to voice it for no particular reason).
>>>
>>> Simple. Two OSes for the price of one. You can upgrade
>>> any time from W10 to W11.
>>
>> I would not buy a new computer with Windows 10.
>>
>> Why?
>>
>> Well, upgrading would put the effort on me. If there is any problem, it is me who has to fight and solve it. If it comes with W11, it means the machine has been tested with W11 and it will work.
>>
>> If I see a computer sold with W10, I get suspicious. It may be old hardware, packaged and installed one or two years ago, sold now. If not packaged, designed time ago, they imaged the install and put it again and again.
>>
>> Whether one likes W10 more than 11 is irrelevant.
>>
>
> Some people buy equipment, based on finding it has W10.
>
> That makes it a marketing consideration.
>
> Same as when some of the companies ship "new" computers,
> where the CPU SKU is three releases old. The parts are
> just as worthwhile, but that may allow the retail price
> of the box to be lower.
>
> The computer will handle the upgrade from W10 to W11 for you.
> It'll even attempt the upgrade, when the upgrade won't work,
> then roll the machine back. It's not like you have to
> raise a finger to get this to work. No making ISOs required.

If the upgrade doesn't work, it is I who gets the flak.

--
Cheers,
Carlos E.R.

Paul wrote:
> On 11/12/2023 1:06 PM, Bill Bradshaw wrote:
> One
>> thing I would like to do would be see at what speed and temperature
>> the CPU is operating at under each of these setting. But I have not
>> had time to look for a program that might do this. Can you think of
>> one? I am starting to be patient with the computer.
>>
>> <Bill>
>
> I posted a link to XTU.
>
> # You will be using this for monitoring,
> # not to make adjustments (which might not even be available).
> # The supported processor list at the bottom of the web page,
> # there is an invisible scroll bar and if you scroll down far
> enough, # your CPU is in the supported list.
>
>
> https://www.intel.com/content/www/us/en/download/17881/intel-extreme-tuning-utility-intel-xtu.html
>
> # Pictures of XTU
>
>
> https://www.notebookcheck.net/Intel-Extreme-Tuning-Utility-XTU-Undervolting-Guide.272120.0.html
>
> Paul

So I ran the program Benchmark with the computer setting on Extreme
Performance. So what are these telling me?

P-Core Maximum 4.41 GHz
E-Core Maximum 3.59 Ghz
Maximum Temp 92 C

Pretty neat program. I just need to know what it is telling me. The
computer is setup so I can not adjust overclocking which make sense. I have
read that this processor can be overclocked to 5.5 Mhz but that would be in
desktops where you could have many fans and liquid cooling.
--
<Bill>

Brought to you from Anchorage, Alaska

On 11/14/2023 1:58 PM, Bill Bradshaw wrote:
> Paul wrote:
>> On 11/12/2023 1:06 PM, Bill Bradshaw wrote:
>> One
>>> thing I would like to do would be see at what speed and temperature
>>> the CPU is operating at under each of these setting. But I have not
>>> had time to look for a program that might do this. Can you think of
>>> one? I am starting to be patient with the computer.
>>>
>>> <Bill>
>>
>> I posted a link to XTU.
>>
>> # You will be using this for monitoring,
>> # not to make adjustments (which might not even be available).
>> # The supported processor list at the bottom of the web page,
>> # there is an invisible scroll bar and if you scroll down far
>> enough, # your CPU is in the supported list.
>>
>>
>> https://www.intel.com/content/www/us/en/download/17881/intel-extreme-tuning-utility-intel-xtu.html
>>
>> # Pictures of XTU
>>
>>
>> https://www.notebookcheck.net/Intel-Extreme-Tuning-Utility-XTU-Undervolting-Guide.272120.0.html
>>
>> Paul
>
> So I ran the program Benchmark with the computer setting on Extreme
> Performance. So what are these telling me?
>
> P-Core Maximum 4.41 GHz
> E-Core Maximum 3.59 Ghz
> Maximum Temp 92 C
>
> Pretty neat program. I just need to know what it is telling me. The
> computer is setup so I can not adjust overclocking which make sense. I have
> read that this processor can be overclocked to 5.5 Mhz but that would be in
> desktops where you could have many fans and liquid cooling.
>

The version of the utility shown here is pretty basic.

https://www.intel.com/content/www/us/en/gaming/resources/overclocking-xtu-guide.html

In some of the views, there is a graph, but the graph is not labeled properly.

Your processor will turbo a different amount, depending on the number of cores
engaged in turbo-ing. At one time, highest turbo was only achieved on one core,
which was perfectly useless, since OS activity plus the user activity, spilled
over onto more than one core, and the user never seemed to get the Turbo they
expected.

This later changed, to turbo on 1-core or 2-cores, reaching the "Turbo value on the tin".

For some reason, Intel does not like publishing the turbo table. Anandtech
used to like publishing the turbo table. The difference in the numbers here,
could be down to the settings on the power limiter.

https://www.techspot.com/articles-info/2626/images/2023-02-13-image-j_1100.webp

Bill Spec
P-Core Maximum 4.41 GHz 5.5 GHz
E-Core Maximum 3.59 Ghz 4.0 GHz
Maximum Temp 92 C 99 C

It could be that your power limiter is set to way less than 157W.
I presume they would take it right up to Max Temp, if there
was power headroom to do it. Techspot notes that laptops don't generally
set 157W, and it might be 55W or 75W kind of thing.

In the Techspot review

https://www.techspot.com/review/2626-intel-core-i9-13950hx/

https://www.techspot.com/articles-info/2626/bench/30-p.webp

that is a Cinebench R23 multithread test (max cores), conducted at
specific power-limiter points. They would evaluate a run at 55W and 75W.
Then, plot up the points. When yours runs 24 cores, there would be a bench
number, and that would be a bench collected at whatever your power limiter
claimed to be set to in XTU. Raising the power setting, does not necessarily
help, unless there is cooling headroom, as the clock won't go to 5.5GHz unless
the cooling can manage <99C . Some laptops actually shoot past 99C limit a bit,
but that does not make them heroic (the thermal limit is not a silicon limit,
it is an organic package limit, the substrate the die sits on has a 99C limit.

https://files03.tchspt.com/down/CinebenchR23.2.zip

Name: CinebenchR23.2.zip
Size: 261,956,799 bytes (249 MiB)
SHA256: FFF6D34B8F696DE64A534FC9F0788EC1E9F1BDA880A88AED92DF8B96D1797C42

You could run a Cinebench, and while it is running with max cores,
see if it is getting anywhere near the number in the graph for the
same power level.

*******

But that is way too much work, for a start.

We could try the more simple-minded Windows evaluation of your hardware.

The only purpose of running this, is to see if something is grossly mis-tuned.

For example, if Device Manager said your graphics was run by the
"Microsoft Basic Display Adapter", then the two graphics scores would be low.
You have composite graphics and two GPUs, the Intel GPU drives the display
wiring, the NVidia GPU dumps 3D pixmaps into system memory, the Intel GPU
then displays them, when playing a game. This is Optimus/Bumblebee or similar,
driver concept.

https://www.minitool.com/news/windows-experience-score.html

PS> get-ciminstance win32_winsat

CPUScore : 9.6
D3DScore : 9.9
DiskScore : 8.2 <=== you should get a better number than mine, with an NVMe
GraphicsScore : 8.5 <=== placeholder graphics card, what I could find during shortage
MemoryScore : 9.6
TimeTaken : MostRecentAssessment
WinSATAssessmentState : 1
WinSPRLevel : 8.2
PSComputerName :

If I compare the current system to my ten year old system (6C 12T), it looks like this

CPUScore : 9.6 9.1 16C32T versus 6C12T
D3DScore : 9.9 9.9
DiskScore : 8.2 5.9 SATASSD versus SATAHDD
GraphicsScore : 8.5 9.3 1050Ti versus 1080
MemoryScore : 9.6 9.1 DDR4 versus DDR3
TimeTaken : MRA
WinSATAssessmentState : 1
WinSPRLevel : 8.2 5.9 <=== "pick weakest thing..."
PSComputerName :

I'm resorting to that, to see if I can get you to elaborate on
what is slow.

slow program launch (start loading, takes a long time, like it is using sandbox)
system fails to respond to launch (bug in the past, loader didn't start right away)
sluggish graphics but program fast (wrong graphics driver)
throttled disk (overheated NVMe, reduced rates)
CPU clock stays low (power limiter set too low, heatsink has fallen off)

I was hoping the XTU would have a better display of core clock rates,
so you could see the trends better.

Paul

On 11/14/2023 8:51 AM, Carlos E. R. wrote:

>> Some people buy equipment, based on finding it has W10.
>>
>> The computer will handle the upgrade from W10 to W11 for you.
>> It'll even attempt the upgrade, when the upgrade won't work,
>> then roll the machine back. It's not like you have to
>> raise a finger to get this to work. No making ISOs required.
>
> If the upgrade doesn't work, it is I who gets the flak.

It depends on your service model, whether this program and your
relationship with the customer, will achieve a result.

https://www.grc.com/incontrol.htm

Has a sound effect, turn down volume during usage.

All it does, is set around four registry entries for you.
I used to set three of those, manually.

You can use that, to stop version changes while
you (or the customer), are away from the machine.

For security updates, control is more limited. A
customer, by clicking a button in the Windows Update
settings panel, can keep the updates out until
you arrive (once a month) and do them.

Paul

Paul wrote:
> On 11/14/2023 1:58 PM, Bill Bradshaw wrote:
>> Paul wrote:
>>> On 11/12/2023 1:06 PM, Bill Bradshaw wrote:
>>> One
>>>> thing I would like to do would be see at what speed and temperature
>>>> the CPU is operating at under each of these setting. But I have
>>>> not had time to look for a program that might do this. Can you
>>>> think of one? I am starting to be patient with the computer.
>>>>
>>>> <Bill>
>>>
>>> I posted a link to XTU.
>>>
>>> # You will be using this for monitoring,
>>> # not to make adjustments (which might not even be available).
>>> # The supported processor list at the bottom of the web page,
>>> # there is an invisible scroll bar and if you scroll down far
>>> enough, # your CPU is in the supported list.
>>>
>>>
>>> https://www.intel.com/content/www/us/en/download/17881/intel-extreme-tuning-utility-intel-xtu.html
>>>
>>> # Pictures of XTU
>>>
>>>
>>> https://www.notebookcheck.net/Intel-Extreme-Tuning-Utility-XTU-Undervolting-Guide.272120.0.html
>>>
>>> Paul
>>
>> So I ran the program Benchmark with the computer setting on Extreme
>> Performance. So what are these telling me?
>>
>> P-Core Maximum 4.41 GHz
>> E-Core Maximum 3.59 Ghz
>> Maximum Temp 92 C
>>
>> Pretty neat program. I just need to know what it is telling me. The
>> computer is setup so I can not adjust overclocking which make sense.
>> I have read that this processor can be overclocked to 5.5 Mhz but
>> that would be in desktops where you could have many fans and liquid
>> cooling.
>>
>
> The version of the utility shown here is pretty basic.
>
>
> https://www.intel.com/content/www/us/en/gaming/resources/overclocking-xtu-guide.html
>
> In some of the views, there is a graph, but the graph is not labeled
> properly.
>
> Your processor will turbo a different amount, depending on the number
> of cores
> engaged in turbo-ing. At one time, highest turbo was only achieved on
> one core,
> which was perfectly useless, since OS activity plus the user
> activity, spilled
> over onto more than one core, and the user never seemed to get the
> Turbo they
> expected.
>
> This later changed, to turbo on 1-core or 2-cores, reaching the
> "Turbo value on the tin".
>
> For some reason, Intel does not like publishing the turbo table.
> Anandtech
> used to like publishing the turbo table. The difference in the
> numbers here,
> could be down to the settings on the power limiter.
>
> https://www.techspot.com/articles-info/2626/images/2023-02-13-image-j_1100.webp
>
> Bill Spec
> P-Core Maximum 4.41 GHz 5.5 GHz
> E-Core Maximum 3.59 Ghz 4.0 GHz
> Maximum Temp 92 C 99 C
>
> It could be that your power limiter is set to way less than 157W.
> I presume they would take it right up to Max Temp, if there
> was power headroom to do it. Techspot notes that laptops don't
> generally
> set 157W, and it might be 55W or 75W kind of thing.
>
> In the Techspot review
>
> https://www.techspot.com/review/2626-intel-core-i9-13950hx/
>
> https://www.techspot.com/articles-info/2626/bench/30-p.webp
>
> that is a Cinebench R23 multithread test (max cores), conducted at
> specific power-limiter points. They would evaluate a run at 55W and
> 75W. Then, plot up the points. When yours runs 24 cores, there
> would be a bench number, and that would be a bench collected at
> whatever your power limiter claimed to be set to in XTU. Raising
> the power setting, does not necessarily help, unless there is
> cooling headroom, as the clock won't go to 5.5GHz unless the
> cooling can manage <99C . Some laptops actually shoot past 99C
> limit a bit, but that does not make them heroic (the thermal limit
> is not a silicon limit, it is an organic package limit, the substrate
> the die sits on has a 99C limit.
>
> https://files03.tchspt.com/down/CinebenchR23.2.zip
>
> Name: CinebenchR23.2.zip
> Size: 261,956,799 bytes (249 MiB)
> SHA256:
> FFF6D34B8F696DE64A534FC9F0788EC1E9F1BDA880A88AED92DF8B96D1797C42
>
> You could run a Cinebench, and while it is running with max cores,
> see if it is getting anywhere near the number in the graph for the
> same power level.
>
> *******
>
> But that is way too much work, for a start.
>
> We could try the more simple-minded Windows evaluation of your
> hardware.
>
> The only purpose of running this, is to see if something is grossly
> mis-tuned.
>
> For example, if Device Manager said your graphics was run by the
> "Microsoft Basic Display Adapter", then the two graphics scores would
> be low.
> You have composite graphics and two GPUs, the Intel GPU drives the
> display
> wiring, the NVidia GPU dumps 3D pixmaps into system memory, the Intel
> GPU
> then displays them, when playing a game. This is Optimus/Bumblebee or
> similar,
> driver concept.
>
> https://www.minitool.com/news/windows-experience-score.html
>
>> get-ciminstance win32_winsat
>
> CPUScore : 9.6
> D3DScore : 9.9
> DiskScore : 8.2 <=== you should get a better number than
> mine, with an NVMe GraphicsScore : 8.5 <=== placeholder
> graphics card, what I could find during shortage MemoryScore
> : 9.6
> TimeTaken : MostRecentAssessment
> WinSATAssessmentState : 1
> WinSPRLevel : 8.2
> PSComputerName :
>
> If I compare the current system to my ten year old system (6C 12T),
> it looks like this
>
> CPUScore : 9.6 9.1 16C32T versus 6C12T
> D3DScore : 9.9 9.9
> DiskScore : 8.2 5.9 SATASSD versus SATAHDD
> GraphicsScore : 8.5 9.3 1050Ti versus 1080
> MemoryScore : 9.6 9.1 DDR4 versus DDR3
> TimeTaken : MRA
> WinSATAssessmentState : 1
> WinSPRLevel : 8.2 5.9 <=== "pick weakest thing..."
> PSComputerName :
>
> I'm resorting to that, to see if I can get you to elaborate on
> what is slow.
>
> slow program launch (start loading, takes a long
> time, like it is using sandbox) system fails to respond to launch
> (bug in the past, loader didn't start right away) sluggish graphics
> but program fast (wrong graphics driver) throttled disk
> (overheated NVMe, reduced rates) CPU clock stays low
> (power limiter set too low, heatsink has fallen off)
>
> I was hoping the XTU would have a better display of core clock rates,
> so you could see the trends better.
>
> Paul

I am going to take some time to study your email at more depth. You gave me
a lot of information.

<Bill>

On 2023-11-15 15:29, Paul wrote:
> On 11/14/2023 8:51 AM, Carlos E. R. wrote:
>
>>> Some people buy equipment, based on finding it has W10.
>>>
>>> The computer will handle the upgrade from W10 to W11 for you.
>>> It'll even attempt the upgrade, when the upgrade won't work,
>>> then roll the machine back. It's not like you have to
>>> raise a finger to get this to work. No making ISOs required.
>>
>> If the upgrade doesn't work, it is I who gets the flak.
>
> It depends on your service model, whether this program and your
> relationship with the customer, will achieve a result.
>
> https://www.grc.com/incontrol.htm
>
> Has a sound effect, turn down volume during usage.
>
> All it does, is set around four registry entries for you.
> I used to set three of those, manually.
>
> You can use that, to stop version changes while
> you (or the customer), are away from the machine.
>
> For security updates, control is more limited. A
> customer, by clicking a button in the Windows Update
> settings panel, can keep the updates out until
> you arrive (once a month) and do them.

Nah, the arrangement is more lax. They do their updates and most
maintenance, and cry out when something goes bad :-)

--
Cheers,
Carlos E.R.