bit about transistor cost...

E

Edward Hernandez

Guest
The John Doe troll stated the following in message-id
<sdhn7c$pkp$4@dont-email.me>:

> The troll doesn\'t even know how to format a USENET post...

And the John Doe troll stated the following in message-id
<sg3kr7$qt5$1@dont-email.me>:

The reason Bozo cannot figure out how to get Google to keep from
breaking its lines in inappropriate places is because Bozo is
CLUELESS...

And yet, the clueless John Doe troll has itself posted yet another
incorrectly formatted USENET posting on Thu, 9 Dec 2021 06:17:35 -0000
(UTC) in message-id <sos71v$ujh$7@dont-email.me>.

This posting is a public service announcement for any google groups
readers who happen by to point out that the John Doe troll does not even
follow it\'s own rules that it uses to troll other posters.

A2hWwNvD4v14
 
E

Edward Hernandez

Guest
The John Doe troll stated the following in message-id
<sdhn7c$pkp$4@dont-email.me>:

> The troll doesn\'t even know how to format a USENET post...

And the John Doe troll stated the following in message-id
<sg3kr7$qt5$1@dont-email.me>:

The reason Bozo cannot figure out how to get Google to keep from
breaking its lines in inappropriate places is because Bozo is
CLUELESS...

And yet, the clueless John Doe troll has itself posted yet another
incorrectly formatted USENET posting on Thu, 9 Dec 2021 06:18:55 -0000
(UTC) in message-id <sos74f$ujh$8@dont-email.me>.

This posting is a public service announcement for any google groups
readers who happen by to point out that the John Doe troll does not even
follow it\'s own rules that it uses to troll other posters.

0y+n+Wbukoox
 
B

bitrex

Guest
On 12/7/21 5:32 PM, Rick C wrote:
On Monday, December 6, 2021 at 8:14:43 PM UTC-5, Martin Brown wrote:

Likewise with PC\'s. I\'m in the market for a new one right now but I\'m
not convinced that any of them offer single threaded performance that is
3x better than the ancient i7-3770 I have now. That has always been my
upgrade heuristic (used to be every 3 years). Clock speeds have maxed
out and now they are adding more cores (many of which are idle most of
the time). Performance cores and efficient cores is the new selling
point. It looks on paper like the i5-12600K might just pass this test.

I bought an i5 machine and it was a real dog. I said something to the effect that they ran out of ways to add transistors to improve the speed of CPUs a few years ago and someone listed a number of architectural improvements they\'ve added for a 20-30% boost.

It has been quite some time since you could expect significant speed improvements by adding transistors or faster clock speeds. I think it was the Pentium 4 where the clock rate peaked at about 3 GHz by adding pipeline stages for shorter gate delays. But the cost of pipeline stalls pretty much mitigated that advantage. I believe people could overclock the Pentium 3 to run faster than the 4.

A decade ago you could overclock the crap out of a number of CPUs;
stable clock boosts from 3.2 to 4 GHz and 25% performance gains on e.g.
the AMD Phenom II Thuban-core weren\'t uncommon.

Those days are pretty much over you can\'t get gains like that anymore
from overclocking. Incidentally i believe the Thuban core from circa
2010 was the last core from AMD that was entirely hand-designed, after
that AMD moved to tool-assisted layout.

The follow-up Bulldozer core was known to be a poor performer on single
threaded apps, there was a lawsuit over it.
 
B

bitrex

Guest
On 12/9/21 5:52 PM, bitrex wrote:
On 12/8/21 6:21 AM, Martin Brown wrote:
On 07/12/2021 23:21, Lasse Langwadt Christensen wrote:
tirsdag den 7. december 2021 kl. 23.32.09 UTC+1 skrev
gnuarm.del...@gmail.com:
On Monday, December 6, 2021 at 8:14:43 PM UTC-5, Martin Brown wrote:

Likewise with PC\'s. I\'m in the market for a new one right now but I\'m
not convinced that any of them offer single threaded performance
that is
3x better than the ancient i7-3770 I have now. That has always been my
upgrade heuristic (used to be every 3 years). Clock speeds have maxed
out and now they are adding more cores (many of which are idle most of
the time). Performance cores and efficient cores is the new selling
point. It looks on paper like the i5-12600K might just pass this test.
I bought an i5 machine and it was a real dog. I said something to
the effect that they ran out of ways to add transistors to improve
the speed of CPUs a few years ago and someone listed a number of
architectural improvements they\'ve added for a 20-30% boost.

It has been quite some time since you could expect significant speed
improvements by adding transistors or faster clock speeds. I think
it was the Pentium 4 where the clock rate peaked at about 3 GHz by
adding pipeline stages for shorter gate delays. But the cost of
pipeline stalls pretty much mitigated that advantage. I believe
people could overclock the Pentium 3 to run faster than the 4.

It would have needed serious water cooling to overclock a Pentium 3.
My P3 portable actually damaged the surface finish of a table when
left on power running a particularly heavy simulation overnight. Used
on a lap at full speed it would almost certainly have resulted in
serious burns!

https://www.cpubenchmark.net/compare/Intel-i7-3770-vs-Intel-i9-12900KF-vs-Intel-Pentium-4-3.60GHz/896vs4611vs1079


Therein lies the problem. The stuff I am developing only cares about
single thread performance so by moving from the i7-3770 to the latest
and greatest i9-12900 I get just twice the speed for 4x the power
used. It would be a lot more cost effective to buy another 3770 or
4770 (they are practically giving them away now as desktops have
fallen out of fashion).

Curiously I can see what turns out to be a step backwards in the
i7-3770 from my portable which is an i7-2670QM. The latter can
correctly handle sincos simultaneous evaluation without a pipeline
stall in my algorithm but the go faster 3770 cannot. I had assumed
until now that it was a later chip with a lower number until I just
looked it up.

It seems some of the trick used in the slower clocked low power
portable CPUs either don\'t make it into the desktop CPUs or are
inapplicable.

The i5-12600K looks like it might just be good enough. Improvements in
the pipelining, sincos simultaneous evaluation and SSE extensions for
tough floating point problems might just be enough to push it over 3x.
On paper its floating point performance looks OK.


There are web pages that can grind a Haswell-core Celeron N3060 with a
2.4 Ghz boost clock, from ~5 years ago, on a netbook with 4 gigs RAM and
SSD, plus 100 megabit internet connection to a halt all by themselves,
no other tabs open. Example:

https://owlcation.com/stem/I-Found-A-Pretty-Rock-On-The-Beach-And-Wondered-II

Adblock helps a bunch with that page
 
L

Lasse Langwadt Christensen

Guest
fredag den 10. december 2021 kl. 00.00.57 UTC+1 skrev bitrex:
On 12/9/21 5:52 PM, bitrex wrote:
On 12/8/21 6:21 AM, Martin Brown wrote:
On 07/12/2021 23:21, Lasse Langwadt Christensen wrote:
tirsdag den 7. december 2021 kl. 23.32.09 UTC+1 skrev
gnuarm.del...@gmail.com:
On Monday, December 6, 2021 at 8:14:43 PM UTC-5, Martin Brown wrote:

Likewise with PC\'s. I\'m in the market for a new one right now but I\'m
not convinced that any of them offer single threaded performance
that is
3x better than the ancient i7-3770 I have now. That has always been my
upgrade heuristic (used to be every 3 years). Clock speeds have maxed
out and now they are adding more cores (many of which are idle most of
the time). Performance cores and efficient cores is the new selling
point. It looks on paper like the i5-12600K might just pass this test.
I bought an i5 machine and it was a real dog. I said something to
the effect that they ran out of ways to add transistors to improve
the speed of CPUs a few years ago and someone listed a number of
architectural improvements they\'ve added for a 20-30% boost.

It has been quite some time since you could expect significant speed
improvements by adding transistors or faster clock speeds. I think
it was the Pentium 4 where the clock rate peaked at about 3 GHz by
adding pipeline stages for shorter gate delays. But the cost of
pipeline stalls pretty much mitigated that advantage. I believe
people could overclock the Pentium 3 to run faster than the 4.

It would have needed serious water cooling to overclock a Pentium 3.
My P3 portable actually damaged the surface finish of a table when
left on power running a particularly heavy simulation overnight. Used
on a lap at full speed it would almost certainly have resulted in
serious burns!

https://www.cpubenchmark.net/compare/Intel-i7-3770-vs-Intel-i9-12900KF-vs-Intel-Pentium-4-3.60GHz/896vs4611vs1079


Therein lies the problem. The stuff I am developing only cares about
single thread performance so by moving from the i7-3770 to the latest
and greatest i9-12900 I get just twice the speed for 4x the power
used. It would be a lot more cost effective to buy another 3770 or
4770 (they are practically giving them away now as desktops have
fallen out of fashion).

Curiously I can see what turns out to be a step backwards in the
i7-3770 from my portable which is an i7-2670QM. The latter can
correctly handle sincos simultaneous evaluation without a pipeline
stall in my algorithm but the go faster 3770 cannot. I had assumed
until now that it was a later chip with a lower number until I just
looked it up.

It seems some of the trick used in the slower clocked low power
portable CPUs either don\'t make it into the desktop CPUs or are
inapplicable.

The i5-12600K looks like it might just be good enough. Improvements in
the pipelining, sincos simultaneous evaluation and SSE extensions for
tough floating point problems might just be enough to push it over 3x.
On paper its floating point performance looks OK.


There are web pages that can grind a Haswell-core Celeron N3060 with a
2.4 Ghz boost clock, from ~5 years ago, on a netbook with 4 gigs RAM and
SSD, plus 100 megabit internet connection to a halt all by themselves,
no other tabs open. Example:

https://owlcation.com/stem/I-Found-A-Pretty-Rock-On-The-Beach-And-Wondered-II

Adblock helps a bunch with that page

who in their right mind doesn\'t use an adblocker?
 
B

bitrex

Guest
On 12/6/21 4:34 PM, Martin Brown wrote:
On 06/12/2021 19:04, Dimiter_Popoff wrote:
On 12/6/2021 20:47, John Larkin wrote:
On Mon, 6 Dec 2021 20:36:17 +0200, Dimiter_Popoff <dp@tgi-sci.com
wrote:

On 12/6/2021 19:42, jlarkin@highlandsniptechnology.com wrote:
https://www.fabricatedknowledge.com/p/the-rising-tide-of-semiconductor

I\'ve also heard that the cost of one next-gen euv scanner is well over
$200M, and that the design and mask set for a high-end chip costs a
billion dollars.

We just don\'t need few-nm chips.

Gradually electronics design without having access to a silicon factory
becomes useless, hopefully the process is slow enough so we don\'t see
that in full.
Sort of like nowadays you can somehow master an internal combustion
engine if you have a lathe and a milling machine but you have no chance
to make it comparable to those car makers make, not to speak about
cost.

Some things have got good enough. Hammers, spoons, beds, LED lights,
microwave ovens. Moore\'s Law can\'t go on forever, and is probably at
or in same cases past its practical limit.

We don\'t need 3 nm chips to text and twitter. I can\'t imagine my cell
phone needing to be better hardware.

I would like a *lot* more battery life - the speed is more than adequate
for my needs. I\'d trade slower when idle for longer life.

Likewise with PC\'s. I\'m in the market for a new one right now but I\'m
not convinced that any of them offer single threaded performance that is
3x better than the ancient i7-3770 I have now. That has always been my
upgrade heuristic (used to be every 3 years). Clock speeds have maxed
out and now they are adding more cores (many of which are idle most of
the time). Performance cores and efficient cores is the new selling
point. It looks on paper like the i5-12600K might just pass this test.

I recently lept forward a decade from a circa 2010 AMD FX to a Ryzen
5600X, motherboard + processor + 32 gigs of RAM were about $500
delivered from Amazon.

Benchmarked vs a i7-3770 the Ryzen 5600X clobbers it in just about any
task you care to mention:

<https://www.youtube.com/watch?v=FxfwvT7GH4g>
 
B

bitrex

Guest
On 12/9/21 6:09 PM, Lasse Langwadt Christensen wrote:
fredag den 10. december 2021 kl. 00.00.57 UTC+1 skrev bitrex:
On 12/9/21 5:52 PM, bitrex wrote:
On 12/8/21 6:21 AM, Martin Brown wrote:
On 07/12/2021 23:21, Lasse Langwadt Christensen wrote:
tirsdag den 7. december 2021 kl. 23.32.09 UTC+1 skrev
gnuarm.del...@gmail.com:
On Monday, December 6, 2021 at 8:14:43 PM UTC-5, Martin Brown wrote:

Likewise with PC\'s. I\'m in the market for a new one right now but I\'m
not convinced that any of them offer single threaded performance
that is
3x better than the ancient i7-3770 I have now. That has always been my
upgrade heuristic (used to be every 3 years). Clock speeds have maxed
out and now they are adding more cores (many of which are idle most of
the time). Performance cores and efficient cores is the new selling
point. It looks on paper like the i5-12600K might just pass this test.
I bought an i5 machine and it was a real dog. I said something to
the effect that they ran out of ways to add transistors to improve
the speed of CPUs a few years ago and someone listed a number of
architectural improvements they\'ve added for a 20-30% boost.

It has been quite some time since you could expect significant speed
improvements by adding transistors or faster clock speeds. I think
it was the Pentium 4 where the clock rate peaked at about 3 GHz by
adding pipeline stages for shorter gate delays. But the cost of
pipeline stalls pretty much mitigated that advantage. I believe
people could overclock the Pentium 3 to run faster than the 4.

It would have needed serious water cooling to overclock a Pentium 3.
My P3 portable actually damaged the surface finish of a table when
left on power running a particularly heavy simulation overnight. Used
on a lap at full speed it would almost certainly have resulted in
serious burns!

https://www.cpubenchmark.net/compare/Intel-i7-3770-vs-Intel-i9-12900KF-vs-Intel-Pentium-4-3.60GHz/896vs4611vs1079


Therein lies the problem. The stuff I am developing only cares about
single thread performance so by moving from the i7-3770 to the latest
and greatest i9-12900 I get just twice the speed for 4x the power
used. It would be a lot more cost effective to buy another 3770 or
4770 (they are practically giving them away now as desktops have
fallen out of fashion).

Curiously I can see what turns out to be a step backwards in the
i7-3770 from my portable which is an i7-2670QM. The latter can
correctly handle sincos simultaneous evaluation without a pipeline
stall in my algorithm but the go faster 3770 cannot. I had assumed
until now that it was a later chip with a lower number until I just
looked it up.

It seems some of the trick used in the slower clocked low power
portable CPUs either don\'t make it into the desktop CPUs or are
inapplicable.

The i5-12600K looks like it might just be good enough. Improvements in
the pipelining, sincos simultaneous evaluation and SSE extensions for
tough floating point problems might just be enough to push it over 3x.
On paper its floating point performance looks OK.


There are web pages that can grind a Haswell-core Celeron N3060 with a
2.4 Ghz boost clock, from ~5 years ago, on a netbook with 4 gigs RAM and
SSD, plus 100 megabit internet connection to a halt all by themselves,
no other tabs open. Example:

https://owlcation.com/stem/I-Found-A-Pretty-Rock-On-The-Beach-And-Wondered-II

Adblock helps a bunch with that page


who in their right mind doesn\'t use an adblocker?

I only block sites that are disgustingly filled with pop ups and other
crap like that one, or have misbehaving scripts, I don\'t mind a few ads
otherwise I think content creators deserve to get paid.
 
B

bitrex

Guest
On 12/9/21 6:10 PM, bitrex wrote:
On 12/6/21 4:34 PM, Martin Brown wrote:
On 06/12/2021 19:04, Dimiter_Popoff wrote:
On 12/6/2021 20:47, John Larkin wrote:
On Mon, 6 Dec 2021 20:36:17 +0200, Dimiter_Popoff <dp@tgi-sci.com
wrote:

On 12/6/2021 19:42, jlarkin@highlandsniptechnology.com wrote:
https://www.fabricatedknowledge.com/p/the-rising-tide-of-semiconductor


I\'ve also heard that the cost of one next-gen euv scanner is well
over
$200M, and that the design and mask set for a high-end chip costs a
billion dollars.

We just don\'t need few-nm chips.

Gradually electronics design without having access to a silicon
factory
becomes useless, hopefully the process is slow enough so we don\'t see
that in full.
Sort of like nowadays you can somehow master an internal combustion
engine if you have a lathe and a milling machine but you have no
chance
to make it comparable to those car makers make, not to speak about
cost.

Some things have got good enough. Hammers, spoons, beds, LED lights,
microwave ovens. Moore\'s Law can\'t go on forever, and is probably at
or in same cases past its practical limit.

We don\'t need 3 nm chips to text and twitter. I can\'t imagine my cell
phone needing to be better hardware.

I would like a *lot* more battery life - the speed is more than
adequate for my needs. I\'d trade slower when idle for longer life.

Likewise with PC\'s. I\'m in the market for a new one right now but I\'m
not convinced that any of them offer single threaded performance that
is 3x better than the ancient i7-3770 I have now. That has always been
my upgrade heuristic (used to be every 3 years). Clock speeds have
maxed out and now they are adding more cores (many of which are idle
most of the time). Performance cores and efficient cores is the new
selling point. It looks on paper like the i5-12600K might just pass
this test.

I recently lept forward a decade from a circa 2010 AMD FX to a Ryzen
5600X, motherboard + processor + 32 gigs of RAM were about $500
delivered from Amazon.

Benchmarked vs a i7-3770 the Ryzen 5600X clobbers it in just about any
task you care to mention:

https://www.youtube.com/watch?v=FxfwvT7GH4g

Note that game engines largely tend to be single-threaded.
 
E

Edward Hernandez

Guest
The John Doe troll stated the following in message-id
<sdhn7c$pkp$4@dont-email.me>:

> The troll doesn\'t even know how to format a USENET post...

And the John Doe troll stated the following in message-id
<sg3kr7$qt5$1@dont-email.me>:

The reason Bozo cannot figure out how to get Google to keep from
breaking its lines in inappropriate places is because Bozo is
CLUELESS...

And yet, the clueless John Doe troll has itself posted yet another
incorrectly formatted USENET posting on Thu, 9 Dec 2021 23:01:35 -0000
(UTC) in message-id <sou1se$8mg$11@dont-email.me>.

This posting is a public service announcement for any google groups
readers who happen by to point out that the John Doe troll does not even
follow it\'s own rules that it uses to troll other posters.

KXrVUx3kkTF8
 
R

Rick C

Guest
On Friday, December 10, 2021 at 6:47:21 AM UTC-5, Tom Del Rosso wrote:
Rick C wrote:
On Wednesday, December 8, 2021 at 9:20:50 PM UTC-5, Tom Del Rosso
wrote:
Rick C wrote:
It was never about performance, it was just the number of
transistors doubling every 18 to 24 months.
All of that may be true, but Moore\'s observation was simply about
the trend in the number of transistors on a die. That\'s all.
Yes but his observation was that it doubled every year, or else the
8080 would have had only 250 transistors.

Not sure what you are talking about.

\"Moore\'s law is the observation that the number of transistors in a
dense integrated circuit doubles about every two years. Moore\'s law
is an observation and projection of a historical trend. Rather than a
law of physics, it is an empirical relationship linked to gains from
experience in production.\"

Wikipedia. If you don\'t trust them see what Synopsis says or any of
dozens of other sources.

Maybe you intended to say something you didn\'t actually say?
In 1959 it was 2 transistors. In 1971 it was 2000. That\'s a lot closer
to doubling every year.

Moore made his observation in the mid 60\'s. You are trying to apply that retroactively to the entire period from the date the first two transistors were placed on a single die. That\'s a pointless comparison.


\"Moore\'s law is the observation that the number of transistors in a
dense integrated circuit doubles about every two years.\"
But in the 90\'s I always read it was every 18 months, and in the 70\'s I
always read it was every year.

In the 70\'s they projected a billion transistors in 2000, which would be
consistent with every year, but it took longer because it slowed down.

You can look at the graph of densities and find that while there are deviations from the curve, it has remained pretty accurate over the long term. Trying to analyze any given point on the curve has no point. You also need to consider that most such plots are of the total transistors on a given device, not the density of transistors which is what Moore\'s Law is about. For example, through the 90s, the period you are talking about being \"slow\" the leading chips were the various flavors of the Pentium. Until 2002 when the Itanium came out with around four times more transistors than the current Pentium.

I think if you look at feature sizes or transistor area you will see a more consistent curve through the years.


It would be nice to see a reference to Moore that was printed in the
1970\'s.

Before the Internet... Not so easy to find. I wasn\'t even working in electronics until the end of the 70\'s, so nothing in my reading room about that..

--

Rick C.

-+- Get 1,000 miles of free Supercharging
-+- Tesla referral code - https://ts.la/richard11209
 
T

Tom Del Rosso

Guest
Rick C wrote:
On Friday, December 10, 2021 at 6:47:21 AM UTC-5, Tom Del Rosso wrote:
In 1959 it was 2 transistors. In 1971 it was 2000. That\'s a lot
closer to doubling every year.

Moore made his observation in the mid 60\'s. You are trying to apply
that retroactively to the entire period from the date the first two
transistors were placed on a single die. That\'s a pointless
comparison.

Obviously he made his statement based on the trend that started in 59,
and his observation (\"law\") can\'t be applied to that period?

--
Defund the Thought Police
Andiamo Brandon!
 
R

Rick C

Guest
On Saturday, December 11, 2021 at 12:39:02 AM UTC-5, Tom Del Rosso wrote:
Rick C wrote:
On Friday, December 10, 2021 at 6:47:21 AM UTC-5, Tom Del Rosso wrote:
In 1959 it was 2 transistors. In 1971 it was 2000. That\'s a lot
closer to doubling every year.

Moore made his observation in the mid 60\'s. You are trying to apply
that retroactively to the entire period from the date the first two
transistors were placed on a single die. That\'s a pointless
comparison.
Obviously he made his statement based on the trend that started in 59,
and his observation (\"law\") can\'t be applied to that period?

That\'s YOUR period, not his. He looked at the then current trend with the data he had from Intel. Do you have any data for these years? I can\'t find any.

Like I said previously, Moore\'s law is about transistor density. Counting two transistors on a die as the first attempt to integrate circuits is a bit disingenuous. Whatever. It\'s not something worth debating further.

--

Rick C.

-++ Get 1,000 miles of free Supercharging
-++ Tesla referral code - https://ts.la/richard11209
 
J

John Doe

Guest
Rick C <gnuarm.deletethisbit@gmail.com> wrote:

> It\'s not something worth debating further.

Eureka!
 
J

John Doe

Guest
bitrex <user@example.net> wrote:

bitrex wrote:

Benchmarked vs a i7-3770 the Ryzen 5600X clobbers it in just about any
task you care to mention:

https://www.youtube.com/watch?v=FxfwvT7GH4g

Note that game engines largely tend to be single-threaded.

Did you found anything about your follow-up interesting...
 
A

Anthony William Sloman

Guest
On Sunday, December 12, 2021 at 5:29:46 AM UTC+11, John Doe wrote:
Rick C <gnuarm.del...@gmail.com> wrote:

It\'s not something worth debating further.

Eureka!

As if John Doe knew what debating was. His posting style is repeated moronic assertion. Paying any attention to what is being said to him doesn\'t come into it.
It probably couldn\'t since it seems likely that he doesn\'t understand anything but moronic over-simplications (most of them obvious nonsense).

--
Bill Sloman, Sydney
 

Welcome to EDABoard.com

Sponsor

Top