• Competitor rules

    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.

8 core SB anyone?

lol, tbh I didn't even read a review on SB-E as it was seemingly known for AGES what it was. Though a year I ago I had assumed it was an octo core then maybe 6 months ago everyone was only talking about it being a hexcore, which was disappointing, not from a "I might of got it if it had 8 cores" stance. But the sooner we get 8 cores in the high end, the sooner we get 8 cores in the midrange.

Nice to know it really is an octo core, (and that I was right ages ago :p ) and size wise very very nice to know.

If a hex core SB-E was 435mm2, then shrunk to 22nm, you're probably looking at a chip around the size of a 2600k now, IE not a whole lot of room for an extra two cores in Haswell. IF 435mm2 is really 8 cores then that does bode well for Haswell being an 8 core part(something I always assumed but was less certain of in the past few months).

Again I'd point out that, SB-E mature process 435mm2, 2.27billion transistors, 130W and that Bulldozer at 315mm2 on a pretty much brand new 32nm process, at 2billion transistors at 125W is exceptionally GOOD. Yeah, performance per watt isn't there, and flat out performance, that can be improved significantly without a die size increase though. When you factor in that, what 15% maybe of the SB-E core is disabled.

Well actually SB-E power is pretty good as well, AMD's transistor density is pretty god damned awesome though and power when you're talking about 2billion transistors is probably a little BETTER than Intel's 2billion at 125W vs what maybe 1.9-1.95billion active at 130W for Intel.
 
Last edited:
Even the VR-Zone article that softpedia refers to is only speculating about possibility of an 8-core dektop version, which they add will likely cost a lot more. Nowhere, so far, is it actually stated the SB-E will ever have 8 cores, which leaked Intel documentation so far has only stated will be available in the SB-EP chips (ie workstation and server chips). Of course, it will be nice if the speculation actually turn out to be prophetic.

Al this is good news for me - it means I can take my time over the SR-3 dual processor build I have planned, and wait for stepping D before buying the processors :)
 
It is quite sensible stopping at 6 cores for Intel. People who need more cores are, for the most part, using their machines for work so price is a bit less of an issue. If somebody needs 8+ cores the arguement is there for dual xeon. Enthusiasts buying the top end platform are still a minority in the grand scheme of things.
 
It's not really news is it? It doesn't cost Intel any more money to manufacturer an 8 core SB-E then it does a 6 core model so why not give us the super rich an 8 core daddy to play with.


. Enthusiasts buying the top end platform are still a minority in the grand scheme of things.

Yes but is an enthusiast spends £800 to £1000 on a single CPU there's a ton of profit for Intel whereas selling to someone like Dell or HP margins are a lot smaller hence why both Intel and AMD still market this type of stuff to likes to you and me.
 
Yes but is an enthusiast spends £800 to £1000 on a single CPU there's a ton of profit for Intel whereas selling to someone like Dell or HP margins are a lot smaller hence why both Intel and AMD still market this type of stuff to likes to you and me.

I agree Intel make a load of money from the enthusiast market but with the lack of competition we will buy the best available. Sure, we might complain about a lack of 8 core products but Intel aren't losing anything except maybe a bit of goodwill from the PC community and bragging rights.

8 Core desktop chips have the potential to erode a market segment that Intel have even higher margins on. Just look at the cost of Xeons vs i7s and the associated chipsets to oems.

Intel are incredibly astute when it comes to defining their markets. I'm not saying I'm happy with what they are doing, far from it in fact, but I can completely understand their motives.
 
But why do Intel need to make anything with 8 cores yet, or lower their hex core prices?

Theres still no competition against their 2500k other than Intel chips with more cores.

Blame AMD.
 
The way i`m kinda looking at it atm is that, upgrade to SB-E now, ignore the quad ivybridge release, and look forward to IvyBridge-E this time next year. The SB-E chips two of the eight cores and 5 or so meg L3 disabled - we can only assume this`ll get unlocked for the XEON chips only or come into play at a later date..

It`s crazy the premium on the Xeon chips though, as far as i understand you`re paying quadruple the price for multi-cpu support and larger cache. And at that they don`t OC as high as far as i can see..

[edit] I think SB-E is pointless for gaming, quad core Ivybridge chips will probably be a more realistic target for that, more efficient better overclocking etc without paying double hopefully for an extra couple of cores..
 
It`s crazy the premium on the Xeon chips though, as far as i understand you`re paying quadruple the price for multi-cpu support and larger cache. And at that they don`t OC as high as far as i can see..

According to currently available info, the highest performing SB-EP (ie SB-E Xeon) will be an octo-core at 3.1GHz, and will cost less than double the price of the i7-3960X.

Nobody knows how well they will overclock yet. As far as I can tell, the only samples that ahve so far ben released have gone to HPC solution providers, I don't believe anyone that would test it's overclocking capabilities has access to them yet.

Mind you, when I build mine, I'll be happy with 32 cores at 4.5 GHz - moreso than 12 cores at 5 GHz :) And for me, the extra cost is worth it.
 
It is quite sensible stopping at 6 cores for Intel. People who need more cores are, for the most part, using their machines for work so price is a bit less of an issue. If somebody needs 8+ cores the arguement is there for dual xeon. Enthusiasts buying the top end platform are still a minority in the grand scheme of things.

Yet plenty of people buy 6970's, instead of 6870's, or 5770's, loads of people use software that uses 8 threads, a crapload, in fact I'd say the vast majority of home users who current own quad cores.......... do not need a quad core in the slightest. This is life, people buy things they don't need, they buy the fastest thing they can buy, because they want it. They get something faster than they need, because its good or maybe better value, or they just don't know what they are buying.

By your logic, Intel and AMD would never have released quad cores, and wouldn't be releasing 6-8 core chips? Clearly that isn't the case.

But why do Intel need to make anything with 8 cores yet, or lower their hex core prices?

Theres still no competition against their 2500k other than Intel chips with more cores.

Blame AMD.

Seriously, where do people come up with these arguments, it blows my mind how they ignore literally all logic.

Bhavv.... the 2500k IS their competition for one thing. Secondly a x6 and a Bulldozer CAN beat it in several situations, its the very reason they can't beat it often that the X6 is priced cheaply to compete, this is the ONLY reason Intel have the chip at £160, if the X6 was priced at £300, or never came within 50% of the speed of the 2500k, the 2500k would be £300.

THen like I said, even then the 2500k would be their competition.

You have several groups of people some of those include, guys who have 2500k's, and WOULD buy another chip if it was faster and good value...... they aren't buying while that chip isn't there, when it is, they'll buy, extra sale. People who have say a i7/i5 of last gen, they don't see the 2500k as a big step, but a 8 core 22nm chip would be a big step and they won't upgrade till then.

So on and so on, people upgrade when the performance and or cost warrants it, considering dual/quads have been around for what, 6 years or so, there are MILLIONS of people out there who have no clue what IPC is, and think the next upgrade is an 8 core chip, there are millions of people out there who know what IPC is and still won't upgrade till they can get a hex/octo core chip.

This competition talk is completely ridiculous, as is the "but we don't need it yet" talk. History shows us, when AMD or Intel have been ahead, they've added more cores for huge performance boosts. History shows us that, without software that uses it, and without ANY need for it, people buy stuff.

This is plain to see with every single piece of hardware you can buy today...... so how can anyone come up with these "but we don't need it" arguments.

Just imagine it, because Intel were ahead at the time, they just stuck with Q6600 to Q6950's or whatever they were, stayed there for 3 years then only decided to move forwards because AMD caught up? Is there any proof of that happening anywhere in the tech industry, seriously? i7's came out when Intel already had the lead, Sandybridge came out, when Intel already had the lead, the first intel hexcores came out when Intel already had the lead, the second gen hex(but really octo) cores came out when Intel already had the lead...............

Now you're both saying Intel won't move ahead because they already have the lead?

This is ignoring the fact that Piledriver already looks to be a pretty big step up from Bulldozer, and Steamroller(version 3 bulldozer basically, scheduled for 2013) is set to be very very interesting indeed.

Intel still can't compete on the gpu side of their APU, they aren't even remotely close, and Llano is selling every chip it can make there. Piledriver will improve performance across the range Llano serves AND moves into even lower power systems AND higher power systems.

AMD cpu, meh, GPU awesome, Intel cpu awesome, gpu woeful. So Intel have a choice extend the cpu lead, or drastically improve the GPU... something they've failed to do for a decade.

No competition, there isn't anything but competition, on CPU's alone AMD is competiting VERY effectively on price, in APU's AMD are simply better for the systems they are designed for(which is spreading to cover the entire range top to bottom, as Intel are planning on doing also). Intel has its OWN CHIPS to compete with, it has the industry to compete with, and it has customers to compete for. Intel are knee deep in heavy competition everywhere, the only reason they are winning, is because they keep improving....... so your theory is they will stop improving?


As for the reality of 8 core SB-E on desktop, its an almost certainty, anyone who thinks otherwise is mentally deranged. They ARE 8 core chips, you disable parts to bring up yields...... you sell chips for less doing so, SB-E has been plagued with problems for a year, its a yield issue. The question is if they ever fix the yields before its not feasable to bother. IE if Haswell is octo core and launches in June 2013, they won't realease an 8 core SB-E for desktop in February, but if yields are great by Feb 2012, why wouldn't they, it would be insane to sell chips crippled when they didn't need to be.
 
Back
Top Bottom