For you amd guys. Now don't get crazy guys.

[Phynaz]

Wait, this thread is three pages long because somebody managed to overclock a quad to 3.2Ghz?

Somebody check the clock on the forums server, I think it got set back a year.

 

[Idontcare]

Originally posted by: RiverRicer
@Idontcare

The link you provided was interesting, however it seemed rather biased toward the negative regarding issues with the technology. It appears that the article was somewhat older than the posted date in that the Gen 2 Z-RAM had been cooking for some time and AMD already licensed it in Dec. 2006.

This link might be of interest. The one comment received on the article from Sang U. Kim, Ph.D is included. He lost me after the first sentence!

http://www.semiconductor.net/i...icleid=CA6512566#93486

Thanks for the article, I had not seen it.

I agree with the commentor though, the theory is at best described incorrectly from a device physics standpoint and is entirely lacking even the most basic physical evidence to prove it is any more or less functional than the now-defunct Gen1 concept was ever proven out, let alone being anywhere near production ready in any meaningful timeframe.

 

[taltamir]

How much MONEY did that cooling system cost... how much money would it cost to get a higher end C2Q to begin with and OC on air? Also an OC means NOTHING if it is not stable. And if you can't run a superpi for 13 whole seconds straight then it isn't merely unstable, its a complete and utter disaster.

2001-2007 = only ignorance of the market (ie, you think only intel makes CPUs and have NO idea that cheaper/faster stuff exists) would make you buy an intel (instead of AMD).
2007+ = only a total fanboy would buy an AMD. (not ignorance, because everyone that knows of AMD knows of intel... unlike the other case, I know many people who think intel is the only one who makes CPU)

 

[Idontcare]

Originally posted by: taltamir
How much MONEY did that cooling system cost... how much money would it cost to get a higher end C2Q to begin with and OC on air? Also an OC means NOTHING if it is not stable. And if you can't run a superpi for 13 whole seconds straight then it isn't merely unstable, its a complete and utter disaster.

2001-2007 = only ignorance of the market (ie, you think only intel makes CPUs and have NO idea that cheaper/faster stuff exists) would make you buy an intel (instead of AMD).
2007+ = only a total fanboy would buy an AMD. (not ignorance, because everyone that knows of AMD knows of intel... unlike the other case, I know many people who think intel is the only one who makes CPU)

Didn't C2D come out summer 2006?

 

[Dainas]

Even if it took this long, it still feels good to know AMD, in one form or another, is making progress with OC. The 45nm with a much bigger L3 still cannot come too soon though.

 

[Cookie Monster]

Originally posted by: taltamir
How much MONEY did that cooling system cost... how much money would it cost to get a higher end C2Q to begin with and OC on air? Also an OC means NOTHING if it is not stable. And if you can't run a superpi for 13 whole seconds straight then it isn't merely unstable, its a complete and utter disaster.

2001-2007 = only ignorance of the market (ie, you think only intel makes CPUs and have NO idea that cheaper/faster stuff exists) would make you buy an intel (instead of AMD).
2007+ = only a total fanboy would buy an AMD. (not ignorance, because everyone that knows of AMD knows of intel... unlike the other case, I know many people who think intel is the only one who makes CPU)

Well if someones looking for a cheap quad core for his needs at "stock" then theres a reason to go with AMD.

Same goes with Tri Cores and etc. A Tri-Core phenom + 780G or MCP78 would be quite the HTPC not to mention it being very cheap. (Plus when it comes to IGPs, AMD/NV offerings are just miles ahead than the intel counterparts, especially gaming/video apps)

AMD doesn't stand a chance in the enthusiast segment and i think everybody "sane" enough knows by now (hopefully), since they get absolutely murdered by the penryns. But fact is, only a small proportion of consumers OC their CPUs. Most dont even know what a BIOs is. AMD has nothing above $249+ dollar mark to challenge intel with their current lineup anyway. But below $249 they are still competitive for 95% of the potential customers to go with the AMD route. People here always give advice on buying E2x00 and then OCing it 3.0GHz. Yeah for us we understand its the most bang per buck thing to do, but no average joe is going to do that period.

The main point of this thread is that the majority B3s can clock up to 3.0GHz with stock volts which is great news for us as consumers since AMD can release higher clocked versions to be even more competitive at the higher end.

 

[Idontcare]

Originally posted by: Idontcare

Originally posted by: RiverRicer
@Idontcare

The link you provided was interesting, however it seemed rather biased toward the negative regarding issues with the technology. It appears that the article was somewhat older than the posted date in that the Gen 2 Z-RAM had been cooking for some time and AMD already licensed it in Dec. 2006.

This link might be of interest. The one comment received on the article from Sang U. Kim, Ph.D is included. He lost me after the first sentence!

http://www.semiconductor.net/i...icleid=CA6512566#93486

Thanks for the article, I had not seen it.

I agree with the commentor though, the theory is at best described incorrectly from a device physics standpoint and is entirely lacking even the most basic physical evidence to prove it is any more or less functional than the now-defunct Gen1 concept was ever proven out, let alone being anywhere near production ready in any meaningful timeframe.

If you have a future fab subscription (its free for online access) you can see more about Gen2 Zram here:

http://www.future-fab.com/cont...f24_wp_l_gaan_ntds.pdf

They include some experimental results in silicon, so I stand corrected on my statement above.

This is farther than I recall public data being discussed before on the Gen1, so maybe this may see the light of day after all.

 

[v8envy]

Originally posted by: Cookie Monster

Well if someones looking for a cheap quad core for his needs at "stock" then theres a reason to go with AMD.

There is a very very small market for people who need a solution for a highly multithreaded or multiprocessing problem but for whom overall performance is unimportant and yet a $20 total system cost delta is a dealbreaker.

Basically, 'lower end enthusiast mainstream extreme value' customers are it. Everyone else is looking at performance/watt, performance/dollar/watt, or other metric other than #cores/die/dollar.

Everyone else buying a quad is probably not doing it based on need or any rational value analysis. At that point it's marketing vs. customer wallet.

Same goes with Tri Cores and etc. A Tri-Core phenom + 780G or MCP78 would be quite the HTPC not to mention it being very cheap. (Plus when it comes to IGPs, AMD/NV offerings are just miles ahead than the intel counterparts, especially gaming/video apps)

Except that a dual core 2 series is all you need to power an HTPC. A tri-core doesn't offer anything a current dual core doesn't already do. GPU acceleration of multimedia is still a rocky road. If you don't have the media encoded in just the right way with the right codecs and played back with just the right software your IGP does squat. On the other hand, a cpu-only solution works, and cpus faster than about 2.2 ghz dual core can accomplish the task just fine.

The main point of this thread is that the majority B3s can clock up to 3.0GHz with stock volts which is great news for us as consumers since AMD can release higher clocked versions to be even more competitive at the higher end.

Plenty of review sites disagree. Other than newegg customers who are hitting 3.2-3.5 ghz on air on every chip the reviews are it's a slightly better overclocker -- 2.8 ghz rather than 2.6 being the ceiling without going all out. And not on stock volts either.

 

[lopri]

Originally posted by: taltamir
How much MONEY did that cooling system cost... how much money would it cost to get a higher end C2Q to begin with and OC on air? Also an OC means NOTHING if it is not stable. And if you can't run a superpi for 13 whole seconds straight then it isn't merely unstable, its a complete and utter disaster.

2001-2007 = only ignorance of the market (ie, you think only intel makes CPUs and have NO idea that cheaper/faster stuff exists) would make you buy an intel (instead of AMD).
2007+ = only a total fanboy would buy an AMD. (not ignorance, because everyone that knows of AMD knows of intel... unlike the other case, I know many people who think intel is the only one who makes CPU)

I am no fanboy of any sort but I would buy AMD platform. I must confess, that building a computer has been more of a hobby for myself than squeezing the best performance/watt out of given budget. If I can explain this by example - I don't run distributed computing, nor do 3D modeling, nor run a database server. My application usages are roughly - 40% office, 20% internet/virtualization, 20% gaming, 20% multimedia/network in home environment. I do try to expand my interest whenever possible, but whichever way I slice, the truth is that hardware has not been a hindrance to what I do for many years. (A glowing exception would be.. Crysis!)

So I have a huge respect for many fellow forummers who contribute to DC or who run intimidating (to me) applications for living or leisure. I myself, however, find more fun in learning how things work/interact and how I can improve my hardware's performance in my own hands.

In this context, I haven't had much luck with Intel platform. It's kind of two folds (but they're related to each other): First, Intel overclocking has been, to say the least, boring. Second, on many occasions things are not under my control with Intel platform.

When Conroe first came out the performance jump was nothing but impressive. It was refreshing that one can experience vastly superior performance by switching the platform. That excitement, however, lasted about 3 days. There wasn't just much to do for myself other than improving coolings. You get to see the max of your CPU, then you're left with one or two usuable dividers with 2 sticks of memory. Available memory paremeters were limited and those rarely mattered even if they were available. With the FSB subsystem dictating what can be done or what can't be, the only way you'd have some kind of fun would be purchasing a $1K CPU, which I refuse to do.

So basically once you find out your max of your CPU (tracking down FSB walls, etc.) then that's about it. Past that point is largely about trouble-shooting. With the advent of quad-cores, now it's all about voltage calibration to achieve Prime95 stability. Ugh. I like to play with multiplies, dividers, and timings. Not those finiky 2 GTLs that need to be calibrated per core. (the price to pay for dealing with MCM quad.

So yeah, I am ready to try out Phenom platform any day now.