Welcome to the dark ages

[RobertR1]

Intel needs their customer base to cycle out their cpu's and Intel is the one who needs to provide good reasoning for that (in terms of performance). They don't have much choice. The MTFB of systems these days is very long thus you need to get them to cycle out for other reasons that product failure.

The biggest worry would be prices, not innovation. Ofcourse it would be ideal if AMD pulled a magic trick and walked away with performance crown for a few years like back in the FX days but I'm sure Intel has taken many steps to ensure something like that won't happen again.

 

[Munky]

You guys are being too pessimistic, IMO, considering that we don't know exactly what is holding back BD, and what changes Piledriver will bring. And don't forget that AMD has the advantage in the tablet market with Bobcat, and the GPU market, both of which are bound to become increasingly more important in the future. For all you know, the BD flop could pave the way for huge progress at AMD, just like Netburst did for Intel.

 

[exar333]

You guys are being too pessimistic, IMO, considering that we don't know exactly what is holding back BD, and what changes Piledriver will bring. And don't forget that AMD has the advantage in the tablet market with Bobcat, and the GPU market, both of which are bound to become increasingly more important in the future. For all you know, the BD flop could pave the way for huge progress at AMD, just like Netburst did for Intel.

So we just wait until 2016 for the next good product from AMD???

 

[exar333]

Intel needs their customer base to cycle out their cpu's and Intel is the one who needs to provide good reasoning for that (in terms of performance). They don't have much choice. The MTFB of systems these days is very long thus you need to get them to cycle out for other reasons that product failure.

The biggest worry would be prices, not innovation. Ofcourse it would be ideal if AMD pulled a magic trick and walked away with performance crown for a few years like back in the FX days but I'm sure Intel has taken many steps to ensure something like that won't happen again.

The current market just couldn't support $1000 mainstream CPUs these days. There are constraints to how high you can raise prices, even with poor competition.

 

[SickBeast]

You guys are being too pessimistic, IMO, considering that we don't know exactly what is holding back BD, and what changes Piledriver will bring. And don't forget that AMD has the advantage in the tablet market with Bobcat, and the GPU market, both of which are bound to become increasingly more important in the future. For all you know, the BD flop could pave the way for huge progress at AMD, just like Netburst did for Intel.

Fair enough. I just think that we need some sort of breakthrough to end the cycle of minimal progress that we've seen over the past few years. They've got to come up with something at the manufacturing level that will make it possible for us to have 10ghz processors. The cores will always be scalable.

You could well be correct, and AMD may well invent something spectacular in a few years. I'm just getting frustrated with them at this point. I have not been blown away by anything they've come up with since the K8, and that's going back quite some time now.

Time will tell. It's really frustrating waiting for Intel to compete with themselves at this point. It seems silly.

 

[Locut0s]

Fair enough. I just think that we need some sort of breakthrough to end the cycle of minimal progress that we've seen over the past few years. They've got to come up with something at the manufacturing level that will make it possible for us to have 10ghz processors. The cores will always be scalable.

SNIP...

Well two things here. (1) what's holding processor back from running at say 10ghz are fundamental issues with physics. Heat dissipation, electron flow issues, quantum mechanical effects and more. These are not things that are easily innovated around, if at all possible. (2) By and large processors have been "fast enough" for some 90% of what perhaps 90% of the population uses computers for, for quite a while now. Word, Powerpoint, web browsing, and HD Youtube just won't benefit from processors getting any faster. This is why for the last few years the best upgrade you can make to your computer has been a solid state drive. Certainly CPUs will continue to advance, becoming more and more powerful, capable of running more and more tasks at the same time (one of the key reasons to go multicore here) and being more capable of running that 10% of apps faster as well. BUT given that we just aren't going to see some fundamental breakthrough that allows PCs to run at 30ghz or something we are basically stuck with going multi/super/uber core. This also means that a LOT of the potential speed improvements now lie with the software developers as much as with the hardware developers. It doesn't matter much if I have a 20 core processor if programmers can't make their software suitably multithreaded. Breakthroughs in fundamental software programming have and will continue to advance in this regard though.

 

[Dadofamunky]

Fair enough. I just think that we need some sort of breakthrough to end the cycle of minimal progress that we've seen over the past few years. They've got to come up with something at the manufacturing level that will make it possible for us to have 10ghz processors. The cores will always be scalable.

'Minimal progress?" Ahem, those of us who bought Nehalems and SBs would beg to differ with that. I waited a long time to upgrade but was amply rewarded.

I think this is very far from the dark ages when you can open 64-bit Photoshop in two seconds.

 

[Munky]

So we just wait until 2016 for the next good product from AMD???

I don't know where you're getting that number

 

[Locut0s]

'Minimal progress?" Ahem, those of us who bought Nehalems and SBs would beg to differ with that. I waited a long time to upgrade but was amply rewarded.

I think this is very far from the dark ages when you can open 64-bit Photoshop in two seconds.

Are you running this off an SSD? Cause an SSD will do far more for application launch times than any CPU upgrade. I agree with you otherwise though.

 

[Locut0s]

I don't know where you're getting that number

While I don't really agree with him I do think I know roughly where he is getting this from. AMDs road map out to 2014 show a rough 10-15% improvement over BD each year:

If it's true that that's ALL we will be seeing from AMD with no further improvement or surprises at all then that means that by 2014 we will have Excavator only being about 45% faster than DB on average. Problem is Intel is already 40-50% faster than BD now in some not unimportant tasks.

If AMD wants to compete with Intel on the high end then I think it would be better for them to start from scratch with a new design. However I think AMD isn't actually planning to compete with Intel at the high end at all. In which case they may still have a viable business moving forward assuming they don't get pinched too much from the top AND bottom by Intel and ARM.

 

[Soulkeeper]

we've been in a stale market for years imo

every since core2 came out, nothing has really been a worthy upgrade

 

[AtenRa]

While I don't really agree with him I do think I know roughly where he is getting this from. AMDs road map out to 2014 show a rough 10-15% improvement over BD each year:

If it's true that that's ALL we will be seeing from AMD with no further improvement or surprises at all then that means that by 2014 we will have Excavator only being about 45% faster than DB on average. Problem is Intel is already 40-50% faster than BD now in some not unimportant tasks.

If AMD wants to compete with Intel on the high end then I think it would be better for them to start from scratch with a new design. However I think AMD isn't actually planning to compete with Intel at the high end at all. In which case they may still have a viable business moving forward assuming they don't get pinched too much from the top AND bottom by Intel and ARM.

The slide shows performance per watt, we have no idea about the performance increase of Piledriver over current Bulldozer.

Until Intel release more unlocked CPUs or allow an OC through the BCLK i will not recommend Intel CPUs bellow the 2500K.

Athlon 631 and FX4100/6100 when OCed are more than enough to compete and outperform at a lower price all Intel CPUs up to Core i5 2400.

I will put both AMD FX8120 and Intel Core i5 2500K in the same basket, the first will give you better multithreading and 2500K will give you better performance in today's apps.
Between the two IMHO, Platform features, Upgradeability and applications used are the deciding factors and not the CPUs.

 

[Barfo]

Dark Ages? Is RD RAM making a comeback?

 

[Locut0s]

Dark Ages? Is RD RAM making a comeback?

Wow the N64 and PS2 used RDRAM:

http://en.wikipedia.org/wiki/RDRAM

I did not know this.

 

[zlejedi]

Dark Ages? Is RD RAM making a comeback?

Some rumors suggest Rambus ram will be there in high-end radeon 79x0


Also that AMD slide is performance per Watt so it will probably be a tiny bit improvement on IPC and much more agressive clock speed at same TDP sadly.
Phenoms II started with 4 cores at 140W with 3,2 Ghz now they sell 3,7Ghz in 125W tdp iirc ?

 

[Silverforce11]

CPUs haven't mattered in a long long time. It's only now that core 2 duos are limiting in gaming, and only in a few recent games.

If they released IvB with +50% performance over SB, it wouldn't matter for gamers at all when games don't really show the benefit. No, benchmarking at 640 x 480 res doesn't mean shit.

 

[Locut0s]

CPUs haven't mattered in a long long time. It's only now that core 2 duos are limiting in gaming, and only in a few recent games.

If they released IvB with +50% performance over SB, it wouldn't matter for gamers at all when games don't really show the benefit. No, benchmarking at 640 x 480 res doesn't mean shit.

In terms of performance yes, which is why we've basically seen CPUs gradually flat line in terms of peek performance. It's all about performance per watt now, as the future is small, slim and portable.

 

[Texashiker]

While I don't really agree with him I do think I know roughly where he is getting this from. AMDs road map out to 2014 show a rough 10-15% improvement over BD each year:

That 10% - 15% is greatly stagnated from over a decade ago.

 

[Locut0s]

That 10% - 15% is greatly stagnated from over a decade ago.

Indeed. But correct me if I'm wrong weren't CPU cycles a lot longer a decode ago?

 

[zlejedi]

CPUs haven't mattered in a long long time. It's only now that core 2 duos are limiting in gaming, and only in a few recent games.

If they released IvB with +50% performance over SB, it wouldn't matter for gamers at all when games don't really show the benefit. No, benchmarking at 640 x 480 res doesn't mean shit.

While you are free to belive whatever you want please try to avoid spreading your "knowledge" like this since someone by mistake could belive in it

http://www.techspot.com/review/405-the-witcher-2-performance/page8.html

If you are running anything below overclocked previous gen i5 cpu you will be bottlenecked in more demanding games

And it's not only fps you will lose. In games like civ 5 you will be losing several seconds per turn waiting for computer to finish his moves

http://www.purepc.pl/files/Image/artykul_zdjecia/2011/Test_AMD_FX-8150/wykresy/wykres_14.png

 

[exar333]

I don't know where you're getting that number

You said BD could 'pave the way' for AMD like Netburst did for Intel. It took 5 years to ditch Netburst and go back to a better arch. So if AMD wants to go down that path, and have crummy CPUs like Intel did, they are more than welcome to. I was pointing out a poor analogy on your part, and providing a little substance on what going down the NB path could actually look like on a real timeline.

I do agree with the other posters below that 10-15% performance improvements/year on the current AMD plan is non-competitive. ARM will probably overtake them in performance and perf/watt with 5 years if that is their real performance plan. Short your AMD stock now, if you believe this to be true.

 

[Munky]

You said BD could 'pave the way' for AMD like Netburst did for Intel. It took 5 years to ditch Netburst and go back to a better arch. So if AMD wants to go down that path, and have crummy CPUs like Intel did, they are more than welcome to. I was pointing out a poor analogy on your part, and providing a little substance on what going down the NB path could actually look like on a real timeline.

I do agree with the other posters below that 10-15% performance improvements/year on the current AMD plan is non-competitive. ARM will probably overtake them in performance and perf/watt with 5 years if that is their real performance plan. Short your AMD stock now, if you believe this to be true.

That doesn't mean it will take 5 years for AMD to ditch or massively overhaul BD. Moreover, in 5 years, do you know what the CPU landscape will look like? I've already stated that the low power and GPGPU aspects will become increasingly important, and AMD has the advantage in both of those areas. Simply limiting the cpu outlook to high end parts is extremely short-sighted, IMO.

 

[Edrick]

CPUs haven't mattered in a long long time. It's only now that core 2 duos are limiting in gaming, and only in a few recent games. If they released IvB with +50% performance over SB, it wouldn't matter for gamers at all when games don't really show the benefit. No, benchmarking at 640 x 480 res doesn't mean shit.

Too many people here think computers are only used for gaming.