No advancements in cpu tech...
Dan Davis
that there really isn't anything much different out there from when I
bought the thing a year and a half a go.
A bit of researched revealed :
Pentium 4 3.06Ghz released November, 2002
Pentium 4 3.6Ghz current fastest model
That's a 15% increase in speed on the same architecture in 2 years.
AMD seems a bit better off. Not in Mhz, but if I use their PR system
then we've had a 33% increase in 2 years. Still, that's abysmal
compared to the pace of advancement of the prior 15 years.
Dual core is supposedly the next big jump forward. That's scheduled
for Q3 2005; It's likely we won't see them until Q4 2005 is my guess,
and they probably wont be even close to a reasonable price (ie <$500
or so) until 2006.
It looks very much to me like the computer revolution is indeed over.
There's nothing out there that is really compelling as an upgrade in
the last 2 years. I think im going to be sticking with my current
setup for another year, year and a half maybe..
I'm also thinking 2005 will be a really, really bad year for Intel and
AMD.
Dan
Yousuf Khan
> I'm also thinking 2005 will be a really, really bad year for Intel and
> AMD.
I don't know, each is trying new ideas out to sell their processors besides
just clockspeed.
AMD for example is trying to fill out its product line from the bottom to
the top to fill every niche. At the top it's got 64-bit 8-way server chips.
At least a couple of firms are expecting to take that out to 32-way or
higher. At the low-end it's introduced Sempron for mainstream value PC
market. At the super-low-end it's putting Geode embedded into an Internet
appliance.
Intel is trying to add features like Wi-Fi and cheap integrated graphics to
its chipsets. It's also just introduced the HD-Audio standard for multi-way
surround sound from a built-in audio system.
Yousuf Khan
JK
Dan Davis wrote:
> Been looking at something to upgrade my old XP 2500+ to. I'm finding
> that there really isn't anything much different out there from when I
> bought the thing a year and a half a go.
>
> A bit of researched revealed :
>
> Pentium 4 3.06Ghz released November, 2002
> Pentium 4 3.6Ghz current fastest model
>
> That's a 15% increase in speed on the same architecture in 2 years.
>
> AMD seems a bit better off. Not in Mhz, but if I use their PR system
> then we've had a 33% increase in 2 years. Still, that's abysmal
> compared to the pace of advancement of the prior 15 years.
How about 64 bit computing? That is a very major advance.
>
>
> Dual core is supposedly the next big jump forward. That's scheduled
> for Q3 2005;
AMD might release it a bit earlier?
> It's likely we won't see them until Q4 2005 is my guess,
> and they probably wont be even close to a reasonable price (ie <$500
> or so) until 2006.
My guess is that dual core processors from AMD might first be released
around the $1,000 mark.
>
>
> It looks very much to me like the computer revolution is indeed over.
> There's nothing out there that is really compelling as an upgrade in
> the last 2 years.
I disagree. The move to 64 bit computing is huge. We will see the
bulk of it starting in '05 after 64 bit Windows is released.
> I think im going to be sticking with my current
> setup for another year, year and a half maybe..
>
> I'm also thinking 2005 will be a really, really bad year for Intel and
> AMD.
2005 will probably be a great year for AMD. I expect 64 bit computing to
become much more widespread in 2005. Imo 2005 might be a very
bad year for Intel, with huge market share losses.
>
>
> Dan
Dan Davis
> Dan Davis wrote:
> > I'm also thinking 2005 will be a really, really bad year for Intel and
> > AMD.
>
> I don't know, each is trying new ideas out to sell their processors besides
> just clockspeed.
>
I understand that. However, as I posted, these are essentially the
same type of processor - they are both P4s.
> AMD for example is trying to fill out its product line from the bottom to
> the top to fill every niche.
But those niches are defined by what is at the top and what is at the
bottom. The top really hasn't changed much at all.
> At the top it's got 64-bit 8-way server chips.
They've been around a while and their presence doesn't make much
difference when you're talking about desktop PCs. I would point out
that 64-bit 128-way servers have been around at the high end for over
4 years now (HP SuperDome).
> At least a couple of firms are expecting to take that out to 32-way or
> higher. At the low-end it's introduced Sempron for mainstream value PC
> market. At the super-low-end it's putting Geode embedded into an Internet
> appliance.
>
> Intel is trying to add features like Wi-Fi and cheap integrated graphics to
> its chipsets. It's also just introduced the HD-Audio standard for multi-way
> surround sound from a built-in audio system.
>
> Yousuf Khan
I understand what they are doing.
My statement regards the 'why'. The 'why' is obvious - the current
CPUs just aren't significantly faster than the ones from 2 years ago.
The only way to get people to buy is to use these gimmicks. PCIe and
such are real advancements, but the rest of it - mostly stuff the
average person already has, or doesnt need.
-Dan
Judd
"JK" <JK9...@netscape.net> wrote in message
news:41841EA7...@netscape.net...
>
>
> Dan Davis wrote:
>
> > Been looking at something to upgrade my old XP 2500+ to. I'm finding
> > that there really isn't anything much different out there from when I
> > bought the thing a year and a half a go.
> >
> > A bit of researched revealed :
> >
> > Pentium 4 3.06Ghz released November, 2002
> > Pentium 4 3.6Ghz current fastest model
> >
> > That's a 15% increase in speed on the same architecture in 2 years.
> >
> > AMD seems a bit better off. Not in Mhz, but if I use their PR system
> > then we've had a 33% increase in 2 years. Still, that's abysmal
> > compared to the pace of advancement of the prior 15 years.
>
> How about 64 bit computing? That is a very major advance.
It's not an advance. It's been around for years. If it was anything,
Itanium would be the #1 processor. It's nothing, like everything else.
> >
> >
> > Dual core is supposedly the next big jump forward. That's scheduled
> > for Q3 2005;
>
> AMD might release it a bit earlier?
Doubtful
> > It's likely we won't see them until Q4 2005 is my guess,
> > and they probably wont be even close to a reasonable price (ie <$500
> > or so) until 2006.
>
> My guess is that dual core processors from AMD might first be released
> around the $1,000 mark.
>
Perhaps, but who cares. Not many multithreaded apps around and that's even
after hyperthreading made it's way through.
dualcore = dual bore. I'd rather have SMP.
> >
> >
> > It looks very much to me like the computer revolution is indeed over.
> > There's nothing out there that is really compelling as an upgrade in
> > the last 2 years.
>
> I disagree. The move to 64 bit computing is huge. We will see the
> bulk of it starting in '05 after 64 bit Windows is released.
It's not huge and the market has shown that. Keep waiting for 64-bit
windows. It's a ghost as it is. When 64-bit windows comes around, their
will need to be a Visual Studio 64-bit edition to go with it. They may not
release that until 2006. Unless you use, gasp, Intel's compilers, you're
writing assembly. Since most want to use Visual Studio, you won't find a
lot of apps coming out until 2006.
> > I think im going to be sticking with my current
> > setup for another year, year and a half maybe..
> >
> > I'm also thinking 2005 will be a really, really bad year for Intel and
> > AMD.
>
> 2005 will probably be a great year for AMD. I expect 64 bit computing to
> become much more widespread in 2005. Imo 2005 might be a very
> bad year for Intel, with huge market share losses.
Doubt it... Intel will stay the status quo. 64-bit hasn't taken off yet and
they'll have it in all of their product lines (outside of Pentium-M) by
February. Even then, no 64-bit Windows so no huge influx of anything
64-bit. If AMD has a great year, it's because they are extremely price
competitive.
JK
Judd wrote:
> JK = Joke
>
> "JK" <JK9...@netscape.net> wrote in message
> news:41841EA7...@netscape.net...
> >
> >
> > Dan Davis wrote:
> >
> > > Been looking at something to upgrade my old XP 2500+ to. I'm finding
> > > that there really isn't anything much different out there from when I
> > > bought the thing a year and a half a go.
> > >
> > > A bit of researched revealed :
> > >
> > > Pentium 4 3.06Ghz released November, 2002
> > > Pentium 4 3.6Ghz current fastest model
> > >
> > > That's a 15% increase in speed on the same architecture in 2 years.
> > >
> > > AMD seems a bit better off. Not in Mhz, but if I use their PR system
> > > then we've had a 33% increase in 2 years. Still, that's abysmal
> > > compared to the pace of advancement of the prior 15 years.
> >
> > How about 64 bit computing? That is a very major advance.
>
> It's not an advance. It's been around for years.
Not in low priced desktop computers until recently.
> If it was anything,
> Itanium would be the #1 processor. It's nothing, like everything else.
When will we see Itanium based computers for under $1,000?
>
>
> > >
> > >
> > > Dual core is supposedly the next big jump forward. That's scheduled
> > > for Q3 2005;
> >
> > AMD might release it a bit earlier?
>
> Doubtful
We shall see.
>
>
> > > It's likely we won't see them until Q4 2005 is my guess,
> > > and they probably wont be even close to a reasonable price (ie <$500
> > > or so) until 2006.
> >
> > My guess is that dual core processors from AMD might first be released
> > around the $1,000 mark.
> >
>
> Perhaps, but who cares. Not many multithreaded apps around and that's even
> after hyperthreading made it's way through.
> dualcore = dual bore. I'd rather have SMP.
>
> > >
> > >
> > > It looks very much to me like the computer revolution is indeed over.
> > > There's nothing out there that is really compelling as an upgrade in
> > > the last 2 years.
> >
> > I disagree. The move to 64 bit computing is huge. We will see the
> > bulk of it starting in '05 after 64 bit Windows is released.
>
> It's not huge and the market has shown that.
How have they shown that? 64 bit Windows for X86-64 isn't even out yet,
and AMD has no trouble at all selling all the Athlon 64 and Opteron chips
it can make. The vast majority of pc users still use Windows.
> Keep waiting for 64-bit
> windows. It's a ghost as it is. When 64-bit windows comes around, their
> will need to be a Visual Studio 64-bit edition to go with it.
Why? The 32 bit version will run just fine with a 64 bit OS. The 64 bit version
will provide even greater performance. Some things can't even be done at
all with a 32 bit OS and 32 bit applications.
http://www.short-media.com/review.php?r=257&p=1
> They may not
> release that until 2006. Unless you use, gasp, Intel's compilers, you're
> writing assembly. Since most want to use Visual Studio, you won't find a
> lot of apps coming out until 2006.
Some people are running 64 bit Linux on Athlon 64 and Opteron systems,
and writing their own programs and generating 64 bit code.
>
>
> > > I think im going to be sticking with my current
> > > setup for another year, year and a half maybe..
> > >
> > > I'm also thinking 2005 will be a really, really bad year for Intel and
> > > AMD.
> >
> > 2005 will probably be a great year for AMD. I expect 64 bit computing to
> > become much more widespread in 2005. Imo 2005 might be a very
> > bad year for Intel, with huge market share losses.
>
> Doubt it... Intel will stay the status quo. 64-bit hasn't taken off yet and
> they'll have it in all of their product lines (outside of Pentium-M) by
> February.
Intel won't have on chip memory controllers on their 64 bit chips though.
Expect poor 64 bit performance from Intel's 64 bit chips compared to
comparably priced AMD 64 bit chips.
> Even then, no 64-bit Windows so no huge influx of anything
> 64-bit. If AMD has a great year, it's because they are extremely price
> competitive.
AMD always was extremely price competative. Many Opteron processors
are going into supercomputers. Look at the Cray supercomputers with
over 1,000 Opteron chips each for example.
Judd
>
>
> Judd wrote:
>
> > JK = Joke
> >
> > "JK" <JK9...@netscape.net> wrote in message
> > news:41841EA7...@netscape.net...
> > >
> > >
> > > Dan Davis wrote:
> > >
> > > > Been looking at something to upgrade my old XP 2500+ to. I'm
finding
> > > > that there really isn't anything much different out there from when
I
> > > > bought the thing a year and a half a go.
> > > >
> > > > A bit of researched revealed :
> > > >
> > > > Pentium 4 3.06Ghz released November, 2002
> > > > Pentium 4 3.6Ghz current fastest model
> > > >
> > > > That's a 15% increase in speed on the same architecture in 2 years.
> > > >
> > > > AMD seems a bit better off. Not in Mhz, but if I use their PR
system
> > > > then we've had a 33% increase in 2 years. Still, that's abysmal
> > > > compared to the pace of advancement of the prior 15 years.
> > >
> > > How about 64 bit computing? That is a very major advance.
> >
> > It's not an advance. It's been around for years.
>
> Not in low priced desktop computers until recently.
>
> > If it was anything,
> > Itanium would be the #1 processor. It's nothing, like everything else.
>
> When will we see Itanium based computers for under $1,000?
When hell freezes over. People who want a PC under $1k don't care whether
it's 64-bit. That's why 64-bit CPU prices are still quite high.
> > > >
> > > > It looks very much to me like the computer revolution is indeed
over.
> > > > There's nothing out there that is really compelling as an upgrade in
> > > > the last 2 years.
> > >
> > > I disagree. The move to 64 bit computing is huge. We will see the
> > > bulk of it starting in '05 after 64 bit Windows is released.
> >
> > It's not huge and the market has shown that.
>
> How have they shown that? 64 bit Windows for X86-64 isn't even out yet,
> and AMD has no trouble at all selling all the Athlon 64 and Opteron chips
> it can make. The vast majority of pc users still use Windows.
OK, but this contradicts what you say below. The reason the move isn't huge
is the reason you stated. There is no 64-bit commodity OS other than Linux.
> > Keep waiting for 64-bit
> > windows. It's a ghost as it is. When 64-bit windows comes around,
their
> > will need to be a Visual Studio 64-bit edition to go with it.
>
> Why? The 32 bit version will run just fine with a 64 bit OS. The 64 bit
version
> will provide even greater performance. Some things can't even be done at
> all with a 32 bit OS and 32 bit applications.
>
> http://www.short-media.com/review.php?r=257&p=1
Do you understand anything? The 32-bit version produces 32-bit code! Damn,
try again!
> > They may not
> > release that until 2006. Unless you use, gasp, Intel's compilers,
you're
> > writing assembly. Since most want to use Visual Studio, you won't find
a
> > lot of apps coming out until 2006.
>
> Some people are running 64 bit Linux on Athlon 64 and Opteron systems,
> and writing their own programs and generating 64 bit code.
So, you say above that the majority use Windows. You say this thing is
huge, yet all you show is Opteron running on Linux. They're running Xeons
on Linux too. So the hell what!
> >
> >
> > > > I think im going to be sticking with my current
> > > > setup for another year, year and a half maybe..
> > > >
> > > > I'm also thinking 2005 will be a really, really bad year for Intel
and
> > > > AMD.
> > >
> > > 2005 will probably be a great year for AMD. I expect 64 bit computing
to
> > > become much more widespread in 2005. Imo 2005 might be a very
> > > bad year for Intel, with huge market share losses.
> >
> > Doubt it... Intel will stay the status quo. 64-bit hasn't taken off yet
and
> > they'll have it in all of their product lines (outside of Pentium-M) by
> > February.
>
> Intel won't have on chip memory controllers on their 64 bit chips though.
> Expect poor 64 bit performance from Intel's 64 bit chips compared to
> comparably priced AMD 64 bit chips.
>
Things will get interesting when the 600 series arrives and the 3.73 GHz EE
which is based on a different architecture. It will take more advantage of
the improvements in bus speeds and Tom's Hardware has shown that you can run
DDR-II 667 on them.
> > Even then, no 64-bit Windows so no huge influx of anything
> > 64-bit. If AMD has a great year, it's because they are extremely price
> > competitive.
>
> AMD always was extremely price competative. Many Opteron processors
> are going into supercomputers. Look at the Cray supercomputers with
> over 1,000 Opteron chips each for example.
Yet it hasn't done much for their bottom line being that price competitive.
There are supercomputers with Itaniums too... several of them. You're not
really making a good argument.
Benjamin Gawert
> AMD always was extremely price competative. Many Opteron processors
> are going into supercomputers. Look at the Cray supercomputers with
> over 1,000 Opteron chips each for example.
People also made a supercomputer out of Playstation 2s. So what?
Benjamin
Dan Davis
The higher end computers you guys are mentioning are getting more
powerful because they are adding more processors than before. That
says absolutely NOTHING about the processor technology itself. As I
mentioned before, 128+ way servers have been around a long time and
stacking up processors does not indicate an advancement in
architecture.
I'm talking about actually making the underlying technology that makes
an individual CPU do more in less time.
THIS is what I'm talking about :
Nov/ Dec 1998 : K6-2 400 / Pentium 3 400
Nov/Dec 1999 : P3-800 / Athlon 800 (One year)
Nov 20th, 2000 : P4 1.4 / 1.5 Ghz (One year)
Nov 2002 : 3.06 Ghz P4 (Two years)
Nov 2004 : 3.6 Ghz P4 / Athlon64 4000+ (Two years)
LOOK at the numbers, both Mhz *and* architecture.
1998 -> 1999 (ONE year) we had a doubling of CPU Mhz AND a more
advanced architecture release (Athlon).
1999->2000 (ONE year) cpu speed DOUBLED again AND a more advanced
architecture released (P4).
2000->2002 (TWO years) speed DOUBLED.
2002-2004 (TWO years) A new architecture was released (Athlon 64).
Speed went up by 15% on the same architecture P4. If you use AMDs PR
system, the A64 gave us 33%.
I think it's crystal clear what I'm talking about here. Look at the
above and try to make excuses if you like, but its plain to see that
processor advancement has been slowing down since 2000. We went from
year over year doubling of speed, to a 2 year doubling of speed, to a
2 year paltry 33% increase. The fact of the matter is, if you bought
a P4 1.5Ghz back in 2000, you'd still have a pretty decent computer
today, 4 years later. Try saying that with any time period in the
1990s!
And on 64bit -> the primary reason for going from 16 - > 32 - > 64 bit
is address space. Some things are better on 64 bit yes, but I submit
that nothing the average desktop user (or average corporate user for
that matter) will be interested in at this point. 64bit will come
because people will need more than the 4GB of RAM 32bit can directly
access, and doing that with a 32bit processor gives you a small
performance hit.
And if none of that convinces you, try looking at the decrease in
activity on these mobo / CPU newsgroups in the last few years.
There's really nothing out there to get excited about, and it shows in
the inactivity. Thats bad news for Intel *and* AMD.
-Dan
Judd
news:88732bea.04110...@posting.google.com...
> This thread seems to be missing the whole point of my original post.
> The higher end computers you guys are mentioning are getting more
> powerful because they are adding more processors than before. That
> says absolutely NOTHING about the processor technology itself. As I
> mentioned before, 128+ way servers have been around a long time and
> stacking up processors does not indicate an advancement in
> architecture.
>
> I'm talking about actually making the underlying technology that makes
> an individual CPU do more in less time.
Dan, you are preaching to the choir. Moore's law is dead. The simple
formula of decreasing transistor size to improve performance has hit a wall.
They knew that years ago though. That's why they've worked so hard at
adding more features to microprocessors. Now, it's all about the features.
Dual "bore" core, hyperthreading, integrated memory controllers, SSE#x,
extremely large caches, better branch prediction, etc., etc. I agree that
the worst of which is yet to come. Dual core... ah, yes, where we decrease
the speed of the processor to add a sister chip onto the die. Great!
Nothing will run faster except those few mutithreaded applications that you
have that properly take advantage of multiple CPUs/virtual CPUs. Everything
else just might run slower. Great! I'm upgrading my system early next
year, after that, I'll wait a few years to even think about it. Serial ATA,
PCI-E, USB2, Firewire 800, DDR-II, 2 MB caches, 64-bit extensions, 1 GHz
bus, onboard Gbit-E. I'm not sure it's going to get any better than that
for a bit unless you're enthralled with wireless or dual bore. Maybe
software will again become the interesting sector.
Yousuf Khan
> My statement regards the 'why'. The 'why' is obvious - the current
> CPUs just aren't significantly faster than the ones from 2 years ago.
> The only way to get people to buy is to use these gimmicks. PCIe and
> such are real advancements, but the rest of it - mostly stuff the
> average person already has, or doesnt need.
Well, I'd say the addition of 64-bit processing is a significant
departure from what was available just two years ago. The first 64-bit
x86 processors started appearing actually about 1.5 years ago.
On the Intel side, they introduced the Prescott Pentium 4, which has 50%
(30 vs. 20) more pipeline stages than the previous generation P4. They
shrunk the die with the introduction of 90nm fabrication, therefore
doubling the L2 cache size (1MB vs. 512K).
On the AMD side, it's even more remarkable. Besides 64-bit, they added
Hypertransport for direct chip-to-chip communications, embedded RAM
controller for lower latency RAM access. These things are adding much
more to performance than simple Mhz improvements. You can tell by
checking out AMD's performance ratings. Athlon 64 is already rated at
4000+ (equivalent to a Pentium 4 4.0Ghz which will never exist, and the
Athlon is just running at 2.4Ghz).
Mhz was starting to run out of steam both as a relative and absolute
performance metric.
Yousuf Khan
Alex Johnson
> These things are adding much
> more to performance than simple Mhz improvements. You can tell by
> checking out AMD's performance ratings. Athlon 64 is already rated at
> 4000+ (equivalent to a Pentium 4 4.0Ghz which will never exist, and the
> Athlon is just running at 2.4Ghz).
Yousuf, I expected a little more from you. You talk as if PR is a
reliable, respected benchmark. It is nothing but an arbitrary
marketting number. It is not verified by anyone respectable, it is not
required to match performance with a Pentium 4's clock speed, and even
AMD /swears/ it has nothing to do with intel chips. They can keep
raising the number by arbitrary amounts all they want, but it tells you
nothing except that they want it to sell better than the lower PR models.
Alex
--
My words are my own. They represent no other; they belong to no other.
Don't read anything into them or you may be required to compensate me
for violation of copyright. (I do not speak for my employer.)
Caecina
>a P4 1.5Ghz back in 2000, you'd still have a pretty decent computer
>today, 4 years later. Try saying that with any time period in the
>1990s!
This is a very good post by Dan. There's no arguing his assertion.
Currently, I'm on a Pentium 4 1.9 GHz, purchased over two years ago, and it's
running quite nicely, handling essentially everything I throw at it.
mike
> Been looking at something to upgrade my old XP 2500+ to. I'm finding
> that there really isn't anything much different out there from when I
> bought the thing a year and a half a go.
>
> A bit of researched revealed :
>
> Pentium 4 3.06Ghz released November, 2002
> Pentium 4 3.6Ghz current fastest model
>
> That's a 15% increase in speed on the same architecture in 2 years.
>
> AMD seems a bit better off. Not in Mhz, but if I use their PR system
> then we've had a 33% increase in 2 years. Still, that's abysmal
> compared to the pace of advancement of the prior 15 years.
>
>
All the advancements from now on are non cpu, ie 3d GPUs etc..