Gaming Octacore Anytime Soon?

Yes I get your point. What I'm saying is that if intel brought out consumer 6 core chips a lot of gamers would upgrade, they would encourage friends to upgrade and it would result in a significant increase in sales. Not to mention companies that was efficiency but don't need xeons.
 
I built my mom a Athlon X2 7850 computer with 2gb RAM and 120gb hard drive using spare parts, approximately 4 years ago. She tells me she has no need for me to upgrade her computer. Similarly within the same timeframe, I built my brother a Athlon II x4 640 based system with a 4850, and my sister a Phenom II 955 with GTX 275, both running 300 gb Velociraptors. None of them have asked me to upgrade their computer, despite me asking them if they want an upgrade. And it's not just my family either. Most of my relatives are running Phenom II x4 level hardware or older, some even running Core2Duos. In fact, one of them even asked me to try to resurrect a Celeron machine. Not a modern Celeron, but an early Core2Duo era Celeron.

The average person isn't going to care about upgrading a computer. They upgrade when it fails or it can no longer do what they want it to do (i.e. not be able to play a game). As long as it does what they want it to do, they don't really care to upgrade. I've invited my brother to play on my computer in nVsurround. He doesn't care, he just prefers using his computer. And it's not like he's a casual player, he is in FPS clans, and often gets the higher KD ratios. As long as it's good enough to allow him to win at his games, he doesn't care about graphics. The group that constantly wants more is a small group. Why is it the desktop market is shrinking while the mobile market is booming in the consumer space?

And no, it's not a fallacy that Intel refuses to release unlocked 8-core processors. They're doing that to segregate and force people that want high clocked 8-core processors to pray the ultra-high premium for them (Xeons), so they can make more money. It is no error on their part that they didn't release an 8-core i7 this generation. It was a business decision to increase their profits. Additionally, the enthusiast market is tiny. The only reason Intel caters to it is that it influences their position in other markets, because those in the enthusiast market can influence other markets. And the tiny bit of profit it brings from selling leftover Xeon chips.

I agree that Intel places an excessively high value on any processor with more than 6 cores, but these "high-clocked 8-core processors" that they want people to buy carry very middling clock speeds that the end user has no ability to increase. These chips are like Bugatti Veyrons that have 90 mph speed limiters on them. So much wasted potential. The E5-2687W and the more recent E5-2697V2 both run so cool under full load it's crazy. Each of these processors has at LEAST an additional 1.0GHz in them, with 1.5GHz likely possible (although no one outside of Intel knows for sure) with good watercooling. I wouldn't care what Intel wants to charge for these chips and would very happily and gladly buy multiple copies of them if the damn things were unlocked. I think that the price of admission for these chips more than justifies unlocking the top models. An additional 1.5GHz of clock speed obtained by overclocking would turn these chips from merely average single threaded performers into all-around monsters.

If Intel's attitude toward enthusiasts over the last few years is "catering to them", your definition of the term differs greatly from mine. I'd be more inclined to say "flipping the bird" or "insulting" rather than "catering". Weak mainstream performance increases from generation to generation, ever-changing mainstream sockets and a HEDT platform that'll be nearly 3 years old by the time its finally refreshed. Add to this a steady, perpetual diet of Xeon rejects, while at the same time watching Intel hamstring the chips that truly have merit and would otherwise be very attractive (the top bin/workstation Xeons) by locking them down tighter than a frog's arse. Even though there is no danger of remarking of the top bin CPUs, they lock them down hard, simply because they can. Sounds more like contempt, rather than catering, to me...:rolleyes:

I'll wager that as time goes on (as a deliberate part of Intel's strategy or not), each successive generation of CPU will overclock more and more poorly, continuing the decline in CPU overhead that we've already noticed from Sandy Bridge to Haswell. Soon, overclocking will be mostly a futile exercise (with only a 100-200 MHz gain possible), at which point unlocked CPUs will be meaningless. Intel may as well just lock down all chips now and get it over with. After all, we're just a small market, right? What could we do about it? Not a thing. Is Intel being benevolent by continuing to produce an unlocked SKU? I highly doubt it.

I fear that Intel is slowly, but surely, working toward the end of the overclockable computer, at which point enthusiasts will be forced into the mainstream segment. Very little tweaking (if any) will be possible, upgrading will be non-existant and there will be no need for socketed processors any longer (which strangely aligns with rumors about Intel moving to BGA-only CPUs in the future for the mainstream). The only difference in motherboards will be their form factor, with almost everything being integrated onto the CPU itself.

The enthusiast hobby we've known for the last decade and a half is slowly dying and I fear it will cease to exist in any meaningful way before the decade is out...:(

Enjoy it while it lasts...
 
Yes I get your point. What I'm saying is that if intel brought out consumer 6 core chips a lot of gamers would upgrade, they would encourage friends to upgrade and it would result in a significant increase in sales. Not to mention companies that was efficiency but don't need xeons.

Efficiency and overclocking does not go hand in hand. They are polar opposites.

Intel does not need to bring out consumer 6-core desktop chips to encourage people to upgrade. People that upgrade to stay with the latest will upgrade regardless. People that wait generations to upgrade will continue to wait. Meanwhile, their Xeon market is a much more important market to appeal to and protect. If a company needs it, they will cough up the big dollars for the Xeons. They may lose 5 i7 sales at the gain of 1 Xeon sale, but the Xeon can easily net them more than 5 times the profit of an i7 sale.

I'm not saying they shouldn't release affordable 6-cores. I'm just saying it makes no business sense for them to do so.

I agree that Intel places an excessively high value on any processor with more than 6 cores, but these "high-clocked 8-core processors" that they want people to buy carry very middling clock speeds that the end user has no ability to increase. These chips are like Bugatti Veyrons that have 90 mph speed limiters on them. So much wasted potential. The E5-2687W and the more recent E5-2697V2 both run so cool under full load it's crazy. Each of these processors has at LEAST an additional 1.0GHz in them, with 1.5GHz likely possible (although no one outside of Intel knows for sure) with good watercooling. I wouldn't care what Intel wants to charge for these chips and would very happily and gladly buy multiple copies of them if the damn things were unlocked. I think that the price of admission for these chips more than justifies unlocking the top models. An additional 1.5GHz of clock speed obtained by overclocking would turn these chips from merely average single threaded performers into all-around monsters.

If Intel's attitude toward enthusiasts over the last few years is "catering to them", your definition of the term differs greatly from mine. I'd be more inclined to say "flipping the bird" or "insulting" rather than "catering". Weak mainstream performance increases from generation to generation, ever-changing mainstream sockets and a HEDT platform that'll be nearly 3 years old by the time its finally refreshed. Add to this a steady, perpetual diet of Xeon rejects, while at the same time watching Intel hamstring the chips that truly have merit and would otherwise be very attractive (the top bin/workstation Xeons) by locking them down tighter than a frog's arse. Even though there is no danger of remarking of the top bin CPUs, they lock them down hard, simply because they can. Sounds more like contempt, rather than catering, to me...:rolleyes:

I'll wager that as time goes on (as a deliberate part of Intel's strategy or not), each successive generation of CPU will overclock more and more poorly, continuing the decline in CPU overhead that we've already noticed from Sandy Bridge to Haswell. Soon, overclocking will be mostly a futile exercise (with only a 100-200 MHz gain possible), at which point unlocked CPUs will be meaningless. Intel may as well just lock down all chips now and get it over with. After all, we're just a small market, right? What could we do about it? Not a thing. Is Intel being benevolent by continuing to produce an unlocked SKU? I highly doubt it.

I fear that Intel is slowly, but surely, working toward the end of the overclockable computer, at which point enthusiasts will be forced into the mainstream segment. Very little tweaking (if any) will be possible, upgrading will be non-existant and there will be no need for socketed processors any longer (which strangely aligns with rumors about Intel moving to BGA-only CPUs in the future for the mainstream). The only difference in motherboards will be their form factor, with almost everything being integrated onto the CPU itself.

The enthusiast hobby we've known for the last decade and a half is slowly dying and I fear it will cease to exist in any meaningful way before the decade is out...:(

Enjoy it while it lasts...

You are a very small market, and not big enough to risk cannibalizing their Xeon sales over, nor worth their time making a special SKU just for you. The price of them making a special SKU just for you is likely much higher than you can afford to pay. Even if you got every like-minded enthusiast with you, it will still not be enough.

i7s (and nearly every consumer chip ever made) has always been and will continue to be rejected server chips.

What you read as "contempt," to them it's just "good business." You only have two choices in the CPU market, and only one real high performance choice, and Intel knows it.

Additionally, regarding overclocking, Sandy Bridge has always been considered the "fluke." Even Intel has said we will likely never see overclocking like was seen on Sandy Bridge, and they said that well before Ivy Bridge's release.

Sandy Bridge's overclocking was the odd one in that only your luck mattered when overclocking it. It didn't matter what cooling you used as long as you kept it under TJMax. Once it reached its hard frequency limit, that's it. No more overclocking, regardless of how much cooling you throw at it. Ivy Bridge and Haswell, though not as good as overclocking on air, actually did scale with better cooling. So you tell me again, is Intel moving towards locking down or encouraging overclocking? I think they're just staying static. How good the chip overclocks is up to luck, and is not their primary concern when designing the chip. If their design happened to be a good overclocker, so be it. If not, who cares.
 
Last edited:
Efficiency and overclocking does not go hand in hand. They are polar opposites.

Intel does not need to bring out consumer 6-core desktop chips to encourage people to upgrade. People that upgrade to stay with the latest will upgrade regardless. People that wait generations to upgrade will continue to wait. Meanwhile, their Xeon market is a much more important market to appeal to and protect. If a company needs it, they will cough up the big dollars for the Xeons. They may lose 5 i7 sales at the gain of 1 Xeon sale, but the Xeon can easily net them more than 5 times the profit of an i7 sale.

I'm not saying they shouldn't release affordable 6-cores. I'm just saying it makes no business sense for them to do so.



You are a very small market, and not big enough to risk cannibalizing their Xeon sales over, nor worth their time making a special SKU just for you. The price of them making a special SKU just for you is likely much higher than you can afford to pay. Even if you got every like-minded enthusiast with you, it will still not be enough.

i7s (and nearly every consumer chip ever made) has always been and will continue to be rejected server chips.

What you read as "contempt," to them it's just "good business." You only have two choices in the CPU market, and only one real high performance choice, and Intel knows it.

Additionally, regarding overclocking, Sandy Bridge has always been considered the "fluke." Even Intel has said we will likely never see overclocking like was seen on Sandy Bridge, and they said that well before Ivy Bridge's release.

Sandy Bridge's overclocking was the odd one in that only your luck mattered when overclocking it. It didn't matter what cooling you used as long as you kept it under TJMax. Once it reached its hard frequency limit, that's it. No more overclocking, regardless of how much cooling you throw at it. Ivy Bridge and Haswell, though not as good as overclocking on air, actually did scale with better cooling. So you tell me again, is Intel moving towards locking down or encouraging overclocking? I think they're just staying static. How good the chip overclocks is up to luck, and is not their primary concern when designing the chip. If their design happened to be a good overclocker, so be it. If not, who cares.

I understand the economics regarding the Xeon vs the i7 but really fail to see the issue with unlocking the top bin (SKU) Xeons to allow folks who want to play with the best chips the ability to do so if they're willing to pay the price of admission. It wouldn't cannibalize Xeon sales, rather the opposite, it should enhance them, but I'd agree that it would be wise to only unlock the top SKU of Xeon to protect their sales. The number of people who'd buy the top Xeons with the intent of extracting additional speed out of them would be small, but any additional sales due to this would be pure gravy to Intel. It would cost them nothing to unlock the top bin Xeons, no new SKU would be required and it would allow Intel to sell a few more of them. Maybe only an additional 1000-2000 CPUs total but at the price of the top SKU Xeon, that's a fair bit of free profit for zero additional expense.

I must disagree when it comes to Intel's stance toward overclocking. With the loss of conventional overclocking (BCLK) on all but the K processors, and the needless disabling of certain features from the K processors, combined with the decreasing headroom in newer CPUs, I feel that Intel's stance of overclocking is clear...they're merely tolerating it for the time being. We'll have to wait until Broadwell to see if the decreasing headroom trend continues. The rumors about BGA CPUs seem to align with such a strategy too well to be coincidence, IMHO. Eventually. I feel that the mainstream enthusiast segment will cease to exist and will be folded into the standard mainstream, likely by the end of the decade at the latest. The HEDT side of things may last a bit longer and may be retained solely for its ability to allow profitable disposal of rejected Xeons, not out of any desire of Intel to "cater" to the enthusiast market.
 
Are there any games out there that will really take advantage of 8 cores? From what I have seen in the Xeon market, more the cores, the slower the Clock. I am assuming to keep them within a certian TDP. From what I've seen so far, Clock > Cores. (2 core min.)
 
Are there any games out there that will really take advantage of 8 cores?

I believe only BF4 for now with speculation that more console based games will have to move to 8 cores because of the weak per core performance of the CPU that is used in consoles.

From what I have seen in the Xeon market, more the cores, the slower the Clock. I am assuming to keep them within a certian TDP.

That is a big part of it. Slower stock clocks CPUs with more cores to keep the stock + turbo TDP under a specified limit.
 
June timeframe for a Haswell-E release from the stuff I have read online. (6 and 8 Core CPU's)
 
I understand the economics regarding the Xeon vs the i7 but really fail to see the issue with unlocking the top bin (SKU) Xeons to allow folks who want to play with the best chips the ability to do so if they're willing to pay the price of admission. It wouldn't cannibalize Xeon sales, rather the opposite, it should enhance them, but I'd agree that it would be wise to only unlock the top SKU of Xeon to protect their sales. The number of people who'd buy the top Xeons with the intent of extracting additional speed out of them would be small, but any additional sales due to this would be pure gravy to Intel. It would cost them nothing to unlock the top bin Xeons, no new SKU would be required and it would allow Intel to sell a few more of them. Maybe only an additional 1000-2000 CPUs total but at the price of the top SKU Xeon, that's a fair bit of free profit for zero additional expense.

I must disagree when it comes to Intel's stance toward overclocking. With the loss of conventional overclocking (BCLK) on all but the K processors, and the needless disabling of certain features from the K processors, combined with the decreasing headroom in newer CPUs, I feel that Intel's stance of overclocking is clear...they're merely tolerating it for the time being. We'll have to wait until Broadwell to see if the decreasing headroom trend continues. The rumors about BGA CPUs seem to align with such a strategy too well to be coincidence, IMHO. Eventually. I feel that the mainstream enthusiast segment will cease to exist and will be folded into the standard mainstream, likely by the end of the decade at the latest. The HEDT side of things may last a bit longer and may be retained solely for its ability to allow profitable disposal of rejected Xeons, not out of any desire of Intel to "cater" to the enthusiast market.

The K CPUs don't even have conventional BCLK overclocking. They have strap overclocking, which is completely different from BCLK overclocking.

BCLK overclocking wasn't disabled. It was made impossible by tying it to the PCI-E, SATA, and USB clocks. Now, whether this was done to limit BCLK overclocking, or to reduce the number of components used and improve efficiency, I don't know, you have to ask Intel engineers.

Decreasing headroom I doubt was done intentionally. Additionally, as I said, while air and watercooling headroom may have decreased, exotic cooling headroom has increased as compared to Sandy Bridge.
 
Decreasing headroom I doubt was done intentionally.

I do not believe decreasing headroom was intentional either. I believe it was a function of the 22nm process not being much better than the 32nm process when it comes to high performance.
 
The K CPUs don't even have conventional BCLK overclocking. They have strap overclocking, which is completely different from BCLK overclocking.

BCLK overclocking wasn't disabled. It was made impossible by tying it to the PCI-E, SATA, and USB clocks. Now, whether this was done to limit BCLK overclocking, or to reduce the number of components used and improve efficiency, I don't know, you have to ask Intel engineers.

Decreasing headroom I doubt was done intentionally. Additionally, as I said, while air and watercooling headroom may have decreased, exotic cooling headroom has increased as compared to Sandy Bridge.

When I mentioned BCLK overclocking, I was referring to strap overclocking. I apologize for not being more clear. All non K SKUs have no access to the BCLK straps and are limited to maybe 5Mhz on the bus, which for all intents and purposes, disables BCLK overclocking in any meaningful way, a fact which I'm sure Intel was well aware of. The problem is that the PCI-E clock generator was integrated onto the die, which makes it impossible to lock at 100MHz.

The disabling of strap overclocking was first implemented with the Nehalem Xeon, which used a patrol circuit on the motherboard and chipset to forbid access to any BCLK strap not permitted by the CPU. The CPU itself was still capable when used on a normal X58 motherboard, however. With the E5 Xeon, this circuit was integrated onto the CPU itself, locking out non-permitted straps. Intel extended the use of this lockout to every CPU it makes, save for the K SKUs, the Extreme Editions, and the single processor E5-1600s (with the exception of the E5-1680V2, the status of which is still unknown 6 months after its launch due to its total absence from the market in anything other than a new Mac Pro).
 
Are there any games out there that will really take advantage of 8 cores? From what I have seen in the Xeon market, more the cores, the slower the Clock. I am assuming to keep them within a certian TDP. From what I've seen so far, Clock > Cores. (2 core min.)

Since both PS4 and Xbone seems to be reserving 2 cores for OS then scalling above 6 cores is questionable for next few years.
 
Probably after the dx12 reveal if it allows for easier core scaling?
 
A little dual core 54w i3-4340 running at 3.6 ghz with HT destroys current games, if they made a K version of the i3's it would probably really cut into i5 sales
 
i believe this same argument was made when Quad cores started coming out....

As stated by both sides

More than 4 cores are not needed for gaming right now, only very very few games and even then they don't use it %00 efficiently for those few games.

%98 of the market Intel and AMD sell to, could not care less if they have 2,4 or more cores, cause they simply don't know, as stated, these people just want things to work and move on, these are the same people who couldn't care less how many HP their care has , as long as it's gets them from A to B.

%1 who do care, mainly people like [H] and other enthusiast sites and know, aren't big enough for intel to cater to separately, but they do still offer what they can, like a K series, they know this group has some push with people as we are the ones who often tell people what to buy, which still on only amounts to a small small % of total sales to either company.

%1 who would pay a fortune for a unlocked Xeon, who hang out on say xtremeforums, already paid a fortune for an Extreme Edition 6 core and those same people will run out and buy an 8 core, then a 10 core and so on, these are the same people who run out and buy quad Titans on launch day, it is practically a day job for them to get the new, over priced gear, bench the crap out of it then sell it.

This is how is has always been and it won't change. They won't make a Xeon Extreme Edition, they have no reason to, Xeon is about being stable, cool and fast. If they did, it would just be an i7 Extreme Edition chip.


Yes I get your point. What I'm saying is that if intel brought out consumer 6 core chips a lot of gamers would upgrade, they would encourage friends to upgrade and it would result in a significant increase in sales. Not to mention companies that was efficiency but don't need xeons.

A lot isn't as many as you may think... just go look at Steam CPU stats. Also many companies only buy Xeon's for high end workstations via OEMs, if not they get 3930k's or 4930's 6 core rigs likely.

A little dual core 54w i3-4340 running at 3.6 ghz with HT destroys current games, if they made a K version of the i3's it would probably really cut into i5 sales

It would, why they have not done it, because the i3 would likely end up like the old Celeron's that could be clocked to the moon and back! They already run low power....and only 2 cores...that would be fun to try.. i say Intel should release an i3 K edition, but you make it cost the same as a regular i5 4500 chip.. (not the K series)
 
Since both PS4 and Xbone seems to be reserving 2 cores for OS then scalling above 6 cores is questionable for next few years.

There is also the issue that the console CPUs are a lot less powerfull than anything in a modern PC - both in IPC and pure clockspeeds. So I doubt the consoles will create a push for any PC that has a new-ish quadcore to upgrade.
 
%1 who do care, mainly people like [H] and other enthusiast sites and know, aren't big enough for intel to cater to separately, but they do still offer what they can, like a K series, they know this group has some push with people as we are the ones who often tell people what to buy, which still on only amounts to a small small % of total sales to either company.

Why should Intel even care about any influence enthusiasts have with their families/friends/co-workers? The amount of viable alternatives to recommend is exceeedingly short. Cancelling the K series would save them two SKUs. What would enthusiasts do in retaliation? Not a thing. The premise of unlocking the top SKUs has merit in the mainstream market as well. Not only would Intel have two fewer SKUs to worry about, we'd have a more fully featured chip, as the locked versions of a CPU have things like virtualization fully enabled.

%1 who would pay a fortune for a unlocked Xeon, who hang out on say xtremeforums, already paid a fortune for an Extreme Edition 6 core and those same people will run out and buy an 8 core, then a 10 core and so on, these are the same people who run out and buy quad Titans on launch day, it is practically a day job for them to get the new, over priced gear, bench the crap out of it then sell it.
Whether or not someone hangs out on xtremeforums is irrelevant. You mention that the purchase of an Extreme Edition 6 core costs a "fortune" and deride the folks who purchase bleeding edge hardware for the sole purpose of benchmarking it and selling it. An Extreme Edition is a waste of time when a more fully featured, less neutered Xeon equivalent is available, and to some folks, the cost isn't a concern, nor is the cost of the other high-end hardware required to make a bleeding edge system perform at its maximum potential. It's a hobby, pure and simple. Some people spend more on it and some spend less. Personally, if a new component comes out that provides a meaningful boost in performance, I'll consider it. I don't just buy the latest and greatest if there is little to no improvement, which is why I skipped Ivy Bridge-E. Benchmarking to me is merely a verification that the system is performing at the level that it should be...I don't buy a system for the sole purpose of benchmarking it. I do consider cost, but primarily in the context of how much additional performance is being offered for the increased price and whether or not the product represents a company's best product. If Intel offered a 15-core, dual-capable workstation Xeon SKU (that happened to be fully unlocked, as the single CPU variants already are), you bet I'd be all over it. It would perform better than any other CPU on the market by a wide margin, and when overclocked would be an absolutely insane computational monster. To me, the cost for such a product would be well worth its price and I'd gladly pay a substantial premium for such a chip.

This is how is has always been and it won't change. They won't make a Xeon Extreme Edition, they have no reason to, Xeon is about being stable, cool and fast. If they did, it would just be an i7 Extreme Edition chip.

The reason why the multiplier lock was created by Intel in the first place was to prevent remarking of lower grade CPUs as higher ones. This was to prevent unscrupulous individuals from exploiting consumers by selling them a lower grade SKU as a higher one. As the top SKUs have nothing above them, locking the top SKUs is not necessary to protect profits or consumers. A Xeon Extreme Edition isn't required, nor is it being requested. Simply unlocking the top SKUs is all that's needed, something that is already done with the E5-1600 series of processor, which is also a "server" processor. Not providing the same treatment to the dual CPU versions is just Intel being ignorant. If the top SKUs were unlocked, there would be absolutely no risk or impact, as servers using these chips completely lack the options in the BIOS to overclock. Servers or workstations that use the E5-1600 chips are having no issue and don't care at all about the CPUs being unlocked, as they have no options for overclocking.
 
Last edited:
Back
Top