Soldato
- Joined
- 22 Apr 2011
- Posts
- 3,859
- Location
- Nottingham
Will probably go to Haswell in Q2 next year if it brings significant improvements over SB.
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
Ivy-E = Socket 2011, and therefore no new features. Intel have decided that 0.05% performance improvements from now on is all they need. They will keep creating huge architectural improvements, but they won't bother upping the specs of the CPUs. Why? Because AMD are no where and they rule the market.
SandyBridge physically has room for 8-cores, but Intel won't bother because they don't need to. Mugs will still pay £800-1000 for the 6-core.
Haswell will be the same. Some quad core with 3-5mb cache and a core clock of 3.3-3.4ghz which might OC 50mhz better than sandybridge, and offer no real world performance improvement.
Personally I think Ivybridge is a complete peice of crap. Sure clock-to-clock its what, 5% faster, however it clocks a good 300-350mhz worse than sandybridge on average when it comes to overclocking due to heat/stability issues.
I was impressed a bit by Sandybridge, but in all honesty nothing has really "rocked the boat" or moved the game forward since the Bloomfield release back in Nov 2008..
Will probably go to Haswell in Q2 next year if it brings significant improvements over SB.
Intel must LOVE you. SB will be fine for gaming for ages to come.
That maybe so, but some people do like to keep on top of things and stay current - for some of us our careers depend on it.
If the SB - IB improvement is anything to go by we can assume the following:
1) IB-E won't be much better than SB-E
2) Haswell won't be much better than IB
Why? First ones fairly obvious. SB-IB all over again. I know the 2 are not comparable but it does seem to me that the odds are it'll be a moderate performance gain rather than anything remotely special.
Second one. Well if SB-E and IB-E stay the enthusiast standard, and in a lot of things (namely gaming which many of us here enjoy) the enthusiast chips aren't really any better than the higher end mainstream chips, then as IB-E is likely a small skip forward in tech, Haswell will be too.
As said. Main improvement will be power efficiency.
Sad really. I really hope AMD come out with an amazing CPU (piledriver) to bring some competition back to the table. This monopoly is doing nobody any good.
If the SB - IB improvement is anything to go by we can assume the following:
1) IB-E won't be much better than SB-E
2) Haswell won't be much better than IB
Why? First ones fairly obvious. SB-IB all over again. I know the 2 are not comparable but it does seem to me that the odds are it'll be a moderate performance gain rather than anything remotely special.
Second one. Well if SB-E and IB-E stay the enthusiast standard, and in a lot of things (namely gaming which many of us here enjoy) the enthusiast chips aren't really any better than the higher end mainstream chips, then as IB-E is likely a small skip forward in tech, Haswell will be too.
As said. Main improvement will be power efficiency.
Sad really. I really hope AMD come out with an amazing CPU (piledriver) to bring some competition back to the table. This monopoly is doing nobody any good.
I guess I may well be in a minority, but I for one look forward to dropping in an 8/10/12 core replacement for my i7-3960x next year. Who knows what Intel will do between now and then; they may drop their plans for any further LGA2011 processors.
Hmmm, I'm not so sure. Granted that any performance gain is not likely to be hugely significant, and the smart money would bet on that being the case, but if you consider the things that hold IB back then things could be very different.
Consider things like IB's very tight voltage range, high temps and the issues that are the likely suspects of the high temps (tighter integration, IGP, etc) and then consider the current differences between SB and SB-E and apply them to IB and IB-E then you start to get the impression that IB-E could potentially be a real IB 'unleashed' product, especially if they keep IB-E's TDP in the same area as SB-E's.
If they, once again, dropped the IGP (just like SB-E) then this would have the potential of freeing up a lot more room on the die and aleviating some of the heat issues caused by the extra millions of transistors - but then this could mean additional cores might appear in place of the IGP, and maybe the heat returns to some extent. Otherwise with four or six cores it could revitalise the overclocking potential.
Also, imagine being able to overclock IB-E with BCLK as well as muliplier - that too enhances things.
Now, if you make a very conservative estimation of the likely base performance increase of only a couple of percent, and then on top add in the factors I've described, you have the potential of having a REALLY interesting product - IB without IGP, larger die, more room, less heat, potentially more optional cores, 40-ish PCI-E 3.0 lanes, and an overclock potential that existing IB owners wish they had from their previous SB products.
If the above turns out to be true then I wouldn't give Haswell a second look and would continue to try and drive whatever I can out of my investment in LGA 2011.
AMD are just so far behind they are not even in the same game any more. It's not going to be changing any time soon, not unless someone like Samsung step up and buy them out and inject more R&D cash, as has been rumoured. AMD's priority is not the desktop or workstation market any more, and that is essentially their choice - they have freely admitted to shareholders that they are chasing the mobile space very aggresively and they will do very well there. They are just not prioritising R&D in the desktop space as they just don't have the money to do that as well as the other things they already do well.
It's not a monopoly, the chip makers/designers are out there - AMD, Qualcom, Nvidia, Arm, TI, they all have the ability to do what Intel does, they just choose not to. That does not make a monopoly, that is the market doing what the market feels comfortable doing with each individual company playing to their own strengths and in their own specialist area. These guys will all take money off Intel in other markets or even shut them out completely - ie, mobile segments (phones, tablets, etc), so it's definately not a monopoly.
Simply put, if you are in the market to buy some kind of compute asset for a desktop or server, you would need to have your head examined to buy anything other than Intel. Likewise, if you are looking at a mobile phone, Intel are now pursuing that space, but you would need your head examined to consider anything other than an Arm Cortex licensed product - ie, Tegra, Snapdragon, etc.
Over a 30% market share I believe is the UK standard for what counts as a monopoly due to sheer market dominance. Intel own more than that in desktop processors and so in the desktop market they are effectively a monopoly. The threat of others joining in doesn't make it any less so. Currently Intel's only reason to improve anything by a real amount is to raise the barriers to entry from these potential competitors.
As for what you said. Yes it's all true. However if I owned intel i'd get my R&D years ahead of the rest, release products with minor improvements and then, if someone comes up with something to compete, unleash some of your R&D discoveries and regain dominance. By releasing your best early you simply make that technology known to the rest, and they can shortcut through your R&D costs.
Just like the 2600k rolled over the 980X in most things i anticipate haswell rolling over IVB-E as well. Kinda glad i switched from 1366 to 1155, mainstream gets the new tech first![]()
Well since IB-E is set to launch AFTER Haswell, I can't see that being the case. That would be just crazy.
Well since IB-E is set to launch AFTER Haswell, I can't see that being the case. That would be just crazy.
In situations where the haswell is using the same amount of cores as the IB-E, why wouldn't the haswell beat it, Core for core performance would be in favour of Haswell.