Anybody Remember Conroe?

[Sheep221]

I don't know why everyone is afraid of ARM taking over x86, that's not even possible, ARM can't just beat what Intel and AMD technically together were developing since 1960s. Some people have just too unreal imagination of near future.

While ARM is known name now, they deserve it they were making cellphone CPUs for many years before they came into prominence after arrival of android, iOS and more.

I personally don't like or hate any company in general but it has been well known for a while that Intel got lazy, I don't need them to release new CPUs when they are not ready, but they keep doing that so we are getting not really big steps between generations.
But I like that AMD didn't compromise on quality and is still using fluxless solder, compared to thermal paste on Intels.

And I do remember conroes, they are still some left in use, but I never owned anything that is C2D beside two penryn pentium laptops. I just don't change hardware that often so I was always upgrading after several generations.

 

[frozentundra123456]

Just to clarify - I'm not saying that review samples are of higher quality than retail samples of the same notebook. What I'm saying is that the high-end notebooks and ultrabooks that you see reviewed on Anandtech have far better power draw compared to lower end models because Intel makes sure that only the most power frugal components are used on high-end models.

This is why there is so much variation between laptops and ultrabooks with the same Intel CPU in them. It's not Anandtech's fault btw - they only get sent high-end stuff for reviewing, but it gives a false impression of how the CPU (and often the HD graphics) performs in a bog-standard ultrabook design.

Really, a high end ultrabook has better components than a 400.00 one? That surely has to be a conspiracy of some sort. If amd had the same quality of components, what is keeping them from setting standards of their own? Business is competitive, do you expect Intel to put inferior components in their top of the line product to make amd look better?

 

[Arkaign]

The OP comparison lacks substance once we look at the numbers..

Conroe at the time not only was 65nm while Athlon X2 was 90nm
but it has also double the transistor budget , and that was without
integrated memory controller , all this for what after all was a marginal improvement over the competition , though compared to their own previous offering....

Anyway , hardly a technical prowess...

That is a very poorly written post, but perhaps I can sort of understand what you're trying to say.

'Marginal improvement'? The lowest-end models of the Conroe launch used less power and often outperformed the highest end AMD products of the time. Not only that, but they were less expensive, and this kind of beating continued through idiotic things like the 'Quadfather', which was again utterly destroyed by Kentsfield.

http://www.anandtech.com/show/2045

"AMD still doesn't stand a chance. "

"with the E6600 offering better performance than AMD's FX-62 flagship in the vast majority of benchmarks."

by Anand Lal Shimpi on July 14, 2006 12:00 AM EST

Overclocking both the E6600 vs overclocking AMD's best just opened that can of whoopass from a mere beating to an absolute bloodbath.

If you try to make inaccurate statements to open, such as suggesting that Conroe was a marginal improvement over Athlon X2, well that just tends to decrease the respect that your opinion will receive.

As for Silvermont, anything is possible. With Intel's focus on mobile now in full effect, it's really a question in my mind of when dominance is achieved, not "if".

 

[Abwx]

http://www.anandtech.com/show/2045

"AMD still doesn't stand a chance. "

"with the E6600 offering better performance than AMD's FX-62 flagship in the vast majority of benchmarks."

by Anand Lal Shimpi on July 14, 2006 12:00 AM EST

I know this review that start with 9 sysmark benches...

http://www.anandtech.com/show/2045/8

Of course , with a bapco/intel bench...


But the picture is less entertaining with winstone , another
synthetic bench.

http://www.anandtech.com/show/2045/10

Or with rendering.

http://www.anandtech.com/show/2045/11

All this with 291 millions transistors while the Athlon
had only 154 millions that included the IMC , it s just
a disaster in the transistors/perfs ratio , lucky that
they were at 65nm while AMD was at 90nm , frankly ,
that s far from being a breakthrough technologicaly.

Despite smaller process they werent even more power
efficient overall.

http://www.anandtech.com/show/2045/7

And if it was so good without dedicated optimisation
as is usual , why thoses scores with wprime ?....


Intel----C2D E5200 ----------------- @3.3 Ghz------------------------> 24.910 Sec

AMD----A64 X2 6000+ ------------- @3.3 Ghz-------------------------> 25.984 Sec


AMD----Athlon x2 5200+ ----------- @2.7 Ghz-----------------------------> 29.734 Sec

Intel----C2D E6300 ----------------- @2.8 Ghz------------------------------> 30,984 Sec

http://forums.quebecgeeks.net/viewtopic.php?id=6244

 

[A5]

20% Better IPC and 25% lower load power for the same performance (FX-62 v. E6600) is a pretty big deal. When you have to start making up metrics to make yourself feel better (performance per xtor, lol) you should probably just stop.

 

[Abwx]

20% Better IPC and 25% lower load power for the same performance (FX-62 v. E6600) is a pretty big deal. When you have to start making up metrics to make yourself feel better (performance per xtor, lol) you should probably just stop.

In Wprime the core duo was barely in par with the X2 ,
hence it had serious IPC advantage only with well
optimized softs and still barely when FPU was involved
as in cinebench/3dsmax.

So it was not better in integer , (wprime) , and not
much better in FP tasks , here an exemple , one core IPC.

where is your 20%.?.

 

[Abwx]

25% lower load power for the same performance (FX-62 v. E6600)

lol , lets watch the full picture, a good reminder.

 

[mrmt]

'Marginal improvement'? The lowest-end models of the Conroe launch used less power and often outperformed the highest end AMD products of the time. Not only that, but they were less expensive, and this kind of beating continued through idiotic things like the 'Quadfather', which was again utterly destroyed by Kentsfield.

It's no use Arkaign. Conroe broke AMD's back at the time, it sent AMD's $1000 processors to less than $300 and destroyed AMD operating results in the following quarters. AMD had no answer for Conroe, not in absolute performance, not in performance per watt, not in performance per dollar. The company would be bankrupt if not for the Globalfoundries spin off.

No need to argue anything with facts of this magnitude backing you up.

 

[bononos]

I know this review that start with 9 sysmark benches...

http://www.anandtech.com/show/2045/8
Of course , with a bapco/intel bench...


But the picture is less entertaining with winstone , another
synthetic bench.
http://www.anandtech.com/show/2045/10
Or with rendering.
http://www.anandtech.com/show/2045/11

All this with 291 millions transistors while the Athlon
had only 154 millions that included the IMC , it s just
a disaster in the transistors/perfs ratio , lucky that
they were at 65nm while AMD was at 90nm , frankly ,
that s far from being a breakthrough technologicaly.

Despite smaller process they werent even more power
efficient overall.

http://www.anandtech.com/show/2045/7

And if it was so good without dedicated optimisation
as is usual , why thoses scores with wprime ?....
......

Maybe anandtech's review was early on when Intel chipsets had poorer power consumption.
http://www.pcstats.com/articleview.cfm?articleid=2097&page=3

Conroe's larger 4Mb cache probably gave it the edge especially in benchmarks which weren't very memory bandwidth intensive even as it bumped up the transistor count. I did remember that Conroe's were generally thought to be cooler running and faster than AMD equivalents.

 

[Abwx]

Maybe anandtech's review was early on when Intel chipsets had poorer power consumption.
http://www.pcstats.com/articleview.cfm?articleid=2097&page=3

Conroe's larger 4Mb cache probably gave it the edge especially in benchmarks which weren't very memory bandwidth intensive even as it bumped up the transistor count. I did remember that Conroe's were generally thought to be cooler running and faster than AMD equivalents.

Between what is thought and reality there s quite a gap
that was and still is entertained for marketing reasons but as
soon as professional aware users are the concern the datas
we re used to see start to change radicaly.

I ll stop there to not pollute too much the thread but not
without ending with an Anand quote and an Intel very
professional slide (perhaps the only one) of theses times ,
far from consumers fancy benches.

Intel's own marketing material seems to admit that a Xeon E5472 with 800MHz memory is just as a fast as AMD's quad-core at 2GHz.

http://it.anandtech.com/show/2386/11

 

[Sweepr]

where is your 20%.?.

Right here:

http://www.tomshardware.com/reviews/processor-architecture-benchmark,2974-15.html

http://www.hardware.fr/articles/778-14/moyenne.html

Overall results in a series of different benchmarks from Toms Hardware and Hardware.fr, not your AMD seal of approval handpicked benchmarks.

 

[Abwx]

Right here:

Overall results on a series of benchmarks from Toms Hardware and Hardware.fr, not your AMD seal of approval handpicked benchmarks.

I used anand reviews , so theses are AMD remote controlled
benchs and site..?..

Anyway , poor results considering the process and xtor
budget gaps , A64 X2 even being old was no pentium 4...

 

[Arkaign]

I know this review that start with 9 sysmark benches...

http://www.anandtech.com/show/2045/8

Of course , with a bapco/intel bench...


But the picture is less entertaining with winstone , another
synthetic bench.

http://www.anandtech.com/show/2045/10

Or with rendering.

http://www.anandtech.com/show/2045/11

All this with 291 millions transistors while the Athlon
had only 154 millions that included the IMC , it s just
a disaster in the transistors/perfs ratio , lucky that
they were at 65nm while AMD was at 90nm , frankly ,
that s far from being a breakthrough technologicaly.

Despite smaller process they werent even more power
efficient overall.

http://www.anandtech.com/show/2045/7

And if it was so good without dedicated optimisation
as is usual , why thoses scores with wprime ?....


Intel----C2D E5200 ----------------- @3.3 Ghz------------------------> 24.910 Sec

AMD----A64 X2 6000+ ------------- @3.3 Ghz-------------------------> 25.984 Sec


AMD----Athlon x2 5200+ ----------- @2.7 Ghz-----------------------------> 29.734 Sec

Intel----C2D E6300 ----------------- @2.8 Ghz------------------------------> 30,984 Sec

http://forums.quebecgeeks.net/viewtopic.php?id=6244

Haha, go ahead and cherry-pick the handful of synthetics and very rare situations where the gap isn't massive. It just makes you look a bit more delusional than the average nutter.

90nm vs. 65m? Doesn't really mean anything, after all what did Brisbane 65nm Athlon X2 do? Not much of anything really.

The truth is that Conroe simply annihilated Athlon X2. Kentsfield annihilated 'Quadfather'. And after air OC? Straight murder. I owned an X2-3800 with a good cooler and decent overclock (2.6Ghz). I replaced it with an E6400, and even stock, it was a better performer, after clocking it to 3Ghz the difference was just ridiculous. It also ran cooler than the X2.

Please try to find some more random benches or obscure forum posts to 'enlighten' us all, lol. It's pretty entertaining seeing someone tie themselves into knots trying to tell us the world is flat.

 

[Arkaign]

I used anand reviews , so theses are AMD remote controlled
benchs and site..?..

Anyway , poor results considering the process and xtor
budget gaps , A64 X2 even being old was no pentium 4...

You furiously dug through all of Anand's data, to fish out the least damaging possible data, and ignored the overall truth that out of every 100 apps, 95 of them would run a little to a LOT slower on an X2 vs. a Conroe. Anand's own conclusion was that the Conroe kicked the X2's butt up and down the street.

Transistor budget gap? Move those goalposts son! The truth of the X2 not being able to keep up is surely not fair! Life's not fair!

 

[CHADBOGA]

The thing I most remember about Conroe was the denial by so many fanboys.

It looks like BayTrail is having a similar effect already.

 

[SiliconWars]

Really, a high end ultrabook has better components than a 400.00 one? That surely has to be a conspiracy of some sort. If amd had the same quality of components, what is keeping them from setting standards of their own? Business is competitive, do you expect Intel to put inferior components in their top of the line product to make amd look better?

No but I expect people to be able to tell when performance/power etc is down to the overall system and not just the SoC.

 

[wlee15]

The thing I most remember about Conroe was the denial by so many fanboys.

It looks like BayTrail is having a similar effect already.

Conore was clearly superior on a micro-architectural level compared to K8 with wider decode width, better instruction latencies on most of the key instructions, wider SIMD units, a far more sophisticated branch prediction, and a better memory system. Silvermont in contrast has less decoder width, inferior instruction latencies, smaller SIMD mul units. It's not even a slam dunk that Silvermont will match Krait's or A15's IPC.

 

[CHADBOGA]

Conore was clearly superior on a micro-architectural level compared to K8 with wider decode width, better instruction latencies on most of the key instructions, wider SIMD units, a far more sophisticated branch prediction, and a better memory system. Silvermont in contrast has less decoder width, inferior instruction latencies, smaller SIMD mul units. It's not even a slam dunk that Silvermont will match Krait's or A15's IPC.

The appropriate metric for the segment BayTrail/Silvermont is in, is performance/watt.

 

[wlee15]

The appropriate metric for the segment BayTrail/Silvermont is in, is performance/watt.

I don't see it haven't much chance in beating Qualcomm's Snapdragon 800 by enough to displace it's LTE Advanced advantage.

 

[dealcorn]

I don't see it haven't much chance in beating Qualcomm's Snapdragon 800 by enough to displace it's LTE Advanced advantage.

Some clarification of your position may be helpful. At Computex 2013 Intel demonstrated streaming 4K video on Baytrail using what appears to be the new Intel XMM7160 LTE modem that the company announced back at MWC 2013. https://gadgetzoneblog.wordpress.co...ing-1080p-playback-and-4k-streaming-over-lte/ Are you saying streaming 4k video is too low a bar?

 

[mrmt]

You furiously dug through all of Anand's data, to fish out the least damaging possible data, and ignored the overall truth that out of every 100 apps, 95 of them would run a little to a LOT slower on an X2 vs. a Conroe. Anand's own conclusion was that the Conroe kicked the X2's butt up and down the street.

JAFCP (Joint AMD Fanboy Conversion Program) is starting to have effects in the community.

 

[NightDreamer]

Conroe was the best overclocking processor ever.

I remember how advanced it was over the P4, even the slowest 1.86ghz E6300 could trounce a 3+ghz P4.

I had my 1.86 at 3.4. Lapped it too, never ran too hot. Still have pics on Photobucket I think.

That was on a budget 965 chipset too.

 

[A5]

Conroe was the best overclocking processor ever.

I remember how advanced it was over the P4, even the slowest 1.86ghz E6300 could trounce a 3+ghz P4.

I had my 1.86 at 3.4. Lapped it too, never ran too hot. Still have pics on Photobucket I think.

That was on a budget 965 chipset too.

It was good, but it isn't the all-time best. You should check out the old-timers thread talking about the Celeron 300A. 50% OC + dual-socket capable...

 

[Arkaign]

It was good, but it isn't the all-time best. You should check out the old-timers thread talking about the Celeron 300A. 50% OC + dual-socket capable...

1.86 to 3.34 was a lot more than 50% overclock, and it brought the entry C2D to beyond X6800 top-chip levels

Point taken on the dual-socket capability and 'cheapness' of the thing though. I seem to remember the 300A being $180 on launch and it stayed that way for a good while (no AMD competition at that time). It's amazing how expensive building a basic computer used to be compared to today. I remember talking my dad into spending about $1,700 for a floppy drive and tiny ~4MB hard drive upgrade for the old AT.

And don't tell me 300A is old-school, I overclocked my 6Mhz IBM AT to 8Mhz! I was 9, but the computer magazines of the day were truly awesome.

I think the first semi-common 100% overclock I can remember was the 1.6A Pentium 4, albeit not with chipsets available during it's launch. I've seen two with my own eyes that ran at 3.2Ghz on 865 mobos. Northwood was a beast, though it taught a lot of people about electromigration during the SNDS episodes. Some people were truly throwing insane volts at them.

 

[Arachnotronic]

No but I expect people to be able to tell when performance/power etc is down to the overall system and not just the SoC.

I guarantee that if you put an ARM SoC in a poorly designed system that it would have lousy battery life.