[coolaler] Devils Canyon: 4.0 base/ 4.4 turbo @ stock

[moonbogg]

idiots? 😛

Awww come on. Its 5ghz though.

 

[SAAA]

Using SPECint for desktop now ?? 😛
Im sure at the same clocks when OCed the difference will be less than 20% :whiste:

This must be the only benchmark that will show 291% performance difference, im sure they tried hard to find the one benchmark to rule them all, 😛

ps: also completely irrelevant for desktop usage, im sure they could show a nice performance gain going from Core 2 Quad to Devils Canyon by using Cinebench or x264 with AVX but noooo they had to use SPECint 😵

Well they were obviously comparing non overclocked results, at the same speed the difference from sandy would be even less.
But their intention is clearly to show that something happened in those 3-5 years and desktop isn't dying, or at least they still care enough to give more performance to those who use it, not people who could live on with a core-duo or similar.
Btw I'm one of those targeted by the slide, still using a core quad feels so old...

 

[RussianSensation]

Markets are segmented by more than just intended use. If specifications are true, not only would the 5820k be more expensive than DC and require a platform/RAM upgrade for current Haswell users, it will have limited PCI-E lanes for graphics cards.

Based on rumors, the Core i7-5820K will still have 1x16+ 1x8 + 1x4 lanes, which is still more than the 16 offered by 4790K.

--

I have seen many posts now about high DDR4 prices but has anyone seen DDR4 in retail yet to confirm this?

Awww come on. Its 5ghz though.

Even if true, I think i7 5820 @ 4.5ghz will be better over the 3-4 year period. For those who keep their system for a while such as i7 920/i7 860 users, a possible 10% trade-off in max clock speed vs. a 50% increase in core count could be worth it. With DX12, we should get a reduction in draw calls overhead and supposedly games will scale better with more cores, although DX12 is still a solid 1.5 years away. X99 will also allow you to do Ultra M.2 PCIe 3.0 x4 + SLI/Tri-SLI without needing an expensive $280+ Z97 PLX board. Also, since it will support BW-E, next year, one can sell the 5820 chip and get a BW-E that on 14nm could itself overclock to 4.7-4.8ghz.

Have to wait for overclocking results on 5820 vs. 4790K. E8600 was very hyped back in the day but in hindsight, Q6600/6700/9550 all turned out to be better picks for gaming. With PS4/XB1 having 6-available cores and seeing how poorly optimized Watch Dogs was, I can see more PC ports starting to use > 6 cores in the next 2+ years. The huge draw for something like 4790K is much lower pricing ($279 at MC and boards like Asrock Extreme 6 for just $130), and a high base clock for those who don't overclock. With the savings you can pick up a 512GB MX100 SSD or put the funds towards a faster GPU. Pleasantly surprised Intel didn't increase the price from 4770K levels. Sounds like a solid upgrade for 1st and 2nd generation i5/i7 users.

 

[moonbogg]

Good stuff RS. I'm looking forward to 5930k benchmarks. Good times if that sucker OCs well.

 

[toyota]

To be honest I have a 4770k and Im thinking of getting this CPU. Mainly beacause my haswell cant go over 4.4 Ghz and it needs the whole 120v from the wall outlet to do so. Still I want to see how it performs first before taking the swing.

you really have to change to Z97 mobo too to get full potential of the 4790k. it would be beyond silly to change all of that out for what result in a miniscule gain even with a good ocing 4790k. at this point you should at the very minimum wait for Broadwell K. the 4790k is really only for those on older i5 or really old i7 setups that have been putting off getting a 3770k or 4770k.

 

[Grooveriding]

The DDR4 pricing paranoia is nonsense. If a 16GB kit of premium DD3 goes for $250, maybe you'll pay $350 for the same in DDR4. If you're buying HW-E, you're not going to care.

These look like nice chips and amount to delivering factory OCed Haswells at the same price point as the previous 4770K. I don't think those in the market for one of these are even considering HW-E anyways.

 

[OCGuy]

Based on rumors, the Core i7-5820K will still have 1x16+ 1x8 + 1x4 lanes, which is still more than the 16 offered by 4790K.

I have seen many posts now about high DDR4 prices but has anyone seen DDR4 in retail yet to confirm this?

Haswell-E is segmented for a reason. The 5820k is certainly a good choice, but when will it be available?

DDR4 will absolutely be more expensive, and likely slower than current DDR3, until it matures and early adopters pay some R&D tax.

It also has a 700mhz slower base frequency, almost double the TDP, and cannot be dropped into a current platform like 1150. The cost of a X99 board is probably not competitive with a Z87 or Z97 chipset, either.


Frankly, they aren't competing with each other.

 

[OCGuy]

FYI

Buyers will pay a 30 percent or higher premium for DDR4 memory compared to DDR3 next year, though the price differential is expected to decline to around 10 percent in 2015, Howard said. Starting in 2016, DDR4 shipments will outpace DDR3, which will ultimately fade away just like the older DDR2 memory.

http://www.pcworld.com/article/2050260/hefty-price-premium-awaits-early-ddr4-memory-adopters.html

 

[Homeles]

The DDR4 pricing paranoia is nonsense. If a 16GB kit of premium DD3 goes for $250, maybe you'll pay $350 for the same in DDR4. If you're buying HW-E, you're not going to care.

Paranoia? Try common sense.

Call me crazy, but $100 is a lot to me. If I saved up for enough for Haswell-E, I'd definitely care about an added $100.

But why would you spend $100 for something that will add next to zero value for most workloads? Heck, just about everyone has DDR3 at this point, so why not reuse it?

Anyway, I think you posted this in the wrong thread... this one doesn't have the DDR4 discussion.

 

[sxr7171]

So I have a 3570k stable at 4.8GHz all four cores on corsair water. This sounds tempting to me. I picked the 3570 over the 3770 because I heard they O/C better due to lack of HT.

I'm going to assume (hopefully not a poor assumption) that an Intel CPU ships with some headroom. I don't think this is the kind of company that ships out a product that is clocked to its maximum potential. If they were they would have shipped a 4GHz Pentium 4.

When reviews come out and those show it can hit 5GHz then this is a no brainer upgrade for me. HT plus improved IPC and I guess a little higher clock speed.

If on the other hand this does not hit 4.8GHz, I'm not going to bother.

 

[Larnz]

Yeah, I'm hoping this gives me a reason to finally upgrade from my Sandybridge 2600-K which has been clocked at 4.6Ghz on just a corsair H50 cooler for what seems like forever, though 3 and a half years is a really good effort. On the other hand the fact that I havent had a reason to upgrade in such a long time is kinda sad as well.

Hopefully the reviews are good as while I like/want to upgrade I can't justify it if there is next to no real benifet for me. Do you think there will be much of a reason to go from a happy stable 2600K @ 4.6Ghz to the 4790K?

 

[RussianSensation]

DDR4 will absolutely be more expensive,

Ok so if a 16GB kit is about $150 for DDR3-1866/2133, with a 30% premium a 16GB DDR4 kit will be $195. That's not a deal breaker.

and likely slower than current DDR3, until it matures and early adopters pay some R&D tax.

DDR4 is already being showcased at 3200mhz and that's no even one of the top RAM brands.

But the X99 platform will have higher memory bandwidth to begin with that compensates for the lower clockspeed. Early DDR4-2133 kits already overclock to 2700+mhz.

With 8 DIMMs on X99 boards, one can also enable RAMDisk and with it a nice performance increase.

It also has a 700mhz slower base frequency, almost double the TDP, and cannot be dropped into a current platform like 1150.

A solid air cooler or an AIO can help you overclock a 6-core processor and deal with the TDP. I don't think power consumption is a factor if the performance is there.

I doubt many people will do a drop-in upgrade for a 4790K on an existing Z87 motherboard either. That's not really a fair comparison either then since X99 will bring USB 3.1 which is 2x faster than 3.0, SATA Express and SATAe. Most people considering 4790K will look to pair with with a Z97 platform, otherwise you are not really comparing similar mobo feature sets.

The cost of a X99 board is probably not competitive with a Z87 or Z97 chipset, either.

That's true but what if the 5820 is a sub-$400 processor as it's rumored to be priced at? 50% more cores for a minimal increase in price offsets the difference in mobo pricing. There are ~ $215 X79 boards with 8 DIMMs, why wouldn't they be $220-230 X99 boards?

Frankly, they aren't competing with each other.

If 4790K can't show consistent overclocking at 4.9-5.0Ghz, then a 4.4.-4.5Ghz 5820 is a strong contender. If it wasn't, people on our boards wouldn't be discussing these 2 options.

Tallying up:

$50-70 premium for 16GB DDR4
$50-70 premium over $150-180 Z97 boards
Assuming sub-$400 CPU, add $60 CPU premium ($339 vs. $399)
= $150-190 premium

This is similar to the premium Q6600/Q9550 had over their higher clocked dual-core parts. We don't know the price of 5820 or how well it or the 4790K overclocks though. $150-190 extra for 50% more cores is not that expensive for those who can use the cores in things other than games.

As you saw with Watch Dogs and VRAM situation, all it takes is just 1-2 games to change the landscape for PC gamers. It will take just 1-2 high profile games like Witcher 3 or the Division or BF5 to take full advantage of 6 cores and it will make the 5820 @ 4.5ghz worth it over 4790K @ 5Ghz.

I wouldn't discount the 5820 just yet.

 

[Idontcare]

So I have a 3570k stable at 4.8GHz all four cores on corsair water. This sounds tempting to me. I picked the 3570 over the 3770 because I heard they O/C better due to lack of HT.

I'm going to assume (hopefully not a poor assumption) that an Intel CPU ships with some headroom. I don't think this is the kind of company that ships out a product that is clocked to its maximum potential. If they were they would have shipped a 4GHz Pentium 4.

When reviews come out and those show it can hit 5GHz then this is a no brainer upgrade for me. HT plus improved IPC and I guess a little higher clock speed.

If on the other hand this does not hit 4.8GHz, I'm not going to bother.

Intel has actually done this in the past (1.13GHz P3 for example), but the people who made those decisions are no longer in the position to make those kinds of decisions anymore.

So what one could argue is that precisely because they have made that mistake in the past, they are hell-bound to not repeat it. And so they are not likely to be launching DC chips that are effectively at their OC limit. These will be cherry-binned chips for sure.

I like how DC is clearly nothing more than what was supposed to come as part and parcel to the Haswell Refresh cycle, but they delayed it and renamed it something cooler for branding purposes.

How much you want to bet that every single Haswell Refresh chip has the exact same "new and improved next-gen TIM" under the IHS?

I'd be absolutely surprised if that wasn't the case.

 

[Skott]

BTW, how long do you guys think it'll be before we start seeing some OC'ed benchmarks?

Give it a few days. Although not much supply on hand and not released until end of month the professional reviewers should start getting the 4790k fairly soon I would imagine. Intel will want it I'm thinking to promote the chip. Some ES have already been given out although its the regular non cherry picked chips I'm more interested in.

 

[RussianSensation]

Paranoia? Try common sense.

Call me crazy, but $100 is a lot to me. If I saved up for enough for Haswell-E, I'd definitely care about an added $100.

But why would you spend $100 for something that will add next to zero value for most workloads? Heck, just about everyone has DDR3 at this point, so why not reuse it?

If you wouldn't pay $100 more to go from 4 to 6 cores, then of course HW-E platform is not for you. In that case, if $80-100 is a lot to you, why would you get an i7 4790K over 4690K?

Hopefully the reviews are good as while I like/want to upgrade I can't justify it if there is next to no real benifet for me. Do you think there will be much of a reason to go from a happy stable 2600K @ 4.6Ghz to the 4790K?

I wouldn't do it at all. Based on your system specs, I'd rather upgrade your videocards and/or get a larger SSD.

 

[moonbogg]

As you saw with Watch Dogs and VRAM situation, all it takes is just 1-2 games to change the landscape for PC gamers. It will take just 1-2 high profile games like Witcher 3 or the Division or BF5 to take full advantage of 6 cores and it will make the 5820 @ 4.5ghz worth it over 4790K @ 5Ghz.

I wouldn't discount the 5820 just yet.

IMO, BF5 is all it would take to make a 6 core chip the go to CPU for many gamers. BF is one of those few games that lots of people just aren't willing to compromise on very much and this franchise generates lots of upgrades. I know for me that BF games are the only thing that has influenced my upgrades since BF2.

 

[StinkyPinky]

Very underwhelming launch to say the least. I had assumed review sites had samples to review.

Now we need to wait another week for the real info to come through.

 

[Homeles]

If you wouldn't pay $100 more to go from 4 to 6 cores, then of course HW-E platform is not for you. In that case, if $80-100 is a lot to you, why would you get an i7 4790K over 4690K?

Huh? I'm not following you. I'm talking about DDR3 vs. DDR4, not about core count.

$100 for 2 more cores would certainly be worth it to me. Heck, even if they weren't hyperthreaded, I'd possibly go for it.

There's no reason for me, or most people here, to spend "$100" extra for DDR4, given the choice. There's just little value in it. DDR3 is ubiquitous, and there undoubtedly won't be any issue getting it at an affordable price later down the road, if you needed to replace or upgrade the RAM. Proof of this? DDR2 is easily obtainable. Fairly certain Crucial still stocks plain old DDR as well, and it's not crazy expensive. The power savings aren't remarkable. Most applications are unaffected by performance. Finally, the higher capacities enabled by DDR4 aren't going to be something most people need either.

Since Haswell-E doesn't have an IGP, the biggest "mainstream" benefit of higher bandwidth isn't applicable. There certainly are benefits to higher bandwidth, particularly for the Haswell-E platform's potential applications, but those aren't my applications. It's not a sin to use Haswell-E if you're not in need of bandwidth -- the "E" platforms are just more capable of solving bandwidth problems.

It seems like most people want DDR4 just for the sake of having a new technology, regardless of its usefulness to them, and are more than happy to fork over "$100" extra for an incremented digit.

 

[frozentundra123456]

I will believe a 400.00 intel six core when I see it, but I think it would sell really well. I do give intel credit for not raising the price on the 4790k though.

 

[crashtech]

How much you want to bet that every single Haswell Refresh chip has the exact same "new and improved next-gen TIM" under the IHS?

I'd be absolutely surprised if that wasn't the case.

Hadn't considered that, but you are probably right. I've built 2 systems with Haswell Refreshes so far, a 4690 and a 4360, and neither of them topped 75°C under stress with stock HSFs. Well, sort of stock. The i5 got a stock i7 HSF, and the i3 got the i5 stock HSF. That system's been working out pretty well since the i7s always get aftermarket cooling...

 

[Grooveriding]

But why would you spend $100 for something that will add next to zero value for most workloads?

Because if you want to run X99 you need DDR4.

$100 is not much considering the total platform upgrade cost and it sounds like it will only possibly be a 30% premium over DDR3, that's peanuts given how cheap DDR3 is.

 

[Homeles]

How much you want to bet that every single Haswell Refresh chip has the exact same "new and improved next-gen TIM" under the IHS?

I'd be absolutely surprised if that wasn't the case.

That's an interesting assertion. The substrate underside on the refresh HSW processors don't have the additional capacitors, and they use different substrates (minor differences on the underside can be seen in the "capacitor rectangle." The way things appear to me is that the new "K" SKUs are on a separate manufacturing line. They could definitely be using the "NGTIM" on the refresh SKUs, but I wager that they aren't.

Edit: Strangely, there's only one review I can find with temperatures, and that's Tweaktown. The temps are considerably lower... but I really do question that site's ability to properly bench hardware. I don't know why you wouldn't publish temperature readings... it's important, particularly for the higher-profile tech sites, that they measure temps. You've got to keep the manufacturers in check, and temperature is something that can be easily manipulated to make it appear like you've got a lead on your competitors, when in reality, you do not.

Because if you want to run X99 you need DDR4.

$100 is not much considering the total platform upgrade cost and it sounds like it will only possibly be a 30% premium over DDR3, that's peanuts given how cheap DDR3 is.

DDR3's actually relatively expensive right now. The oft-cited 30% premium figure was from IHS, and those numbers were polled before the Hynix fire had happened, and the resulting price hike still hasn't subsided.

 

[ehume]

Given the discussion about RAM, I don't feel so bad about asking this side-topic question here:

It looks like DC is going to be good enough to work as a heater for testing heatsinks. My previous heater was an i7 860 that really cooked at 4GHz. An accidental peripheral short killed the MB. (RIP -- sic transit Gloria Mundi.)

But that ran W7-64. In pushing the OC on this, I can still get a W7-64, or move on and get a W8.1-64. Which will OC better? Or does it matter?

 

[blackened23]

Pfft. All of this Haswell-E and DDR4 sidebar. If you think DDR4 is going to be cheap, keep thinking that. Right now 16GB DDR4 kits are going for near 600$. It isn't hard to find prices. Google it. If you think it's going to be remotely reasonable, well, having seen like 5 DRAM launches in prior years. If you've watched prior DRAM launches, you know what's up: they always are overpriced and underperforming at launch. In fact, at launch, every new DRAM technology is priced into the stratosphere at LAUNCH. There has never been an exception. Period. But what will happen is prices will slowly normalize and speeds will increase over the next year or two after launch of the intel platform using it. But, you say, an intel slide told you that prices won't be expensive. Okay. Believe that. They said the same thing about RDRAM. "It wont' have a significant price premium". LOL. Hoooooooooboy did that not happen.

DDR4 will not be an exception. But by all means, keep thinking otherwise. I'll be looking on with amusement when Haswell-E does launch and everyone and their brother is whining about launch DDR 4 prices. You can buy them for servers NOW. 16GB for 600$ of DDR4 at slower rated speeds than comparable DDR3. Have fun with that.

Anyway, Haswell-E will be a great platform for those wanting hexa core or 8 core CPUs. The 5820K should be a great chip. However, make no mistake, DDR4 will be a significant setback for many wanting to go the HEDT route. That will eventually change, but it's going to take at least a year after Haswell-E launches for prices to not be stupid. Most prior DRAM launches took more than a year for prices to go into the sane realm. DDR3 wasn't cheap at launch. Same crap happened. Same crap will happen again.

Then again, does it really freakin matter? Probably not. People adopting the HEDT platform aren't penny pinchers anyway. So in the end it doesn't make a difference. Many guys buying into HEDT area already spending 400$ on a mobo and 600$ on a CPU. Will 500$ for DRAM matter to these guys? Heck no. DDR4 is going to be expensive for the budget minded guys, but most HEDT users simply won't care and will buy it anyway despite the cost.

 

[tential]

Pfft. All of this Haswell-E and DDR4 sidebar. If you think DDR4 is going to be cheap, keep thinking that. Right now 16GB DDR4 kits are going for near 600$. It isn't hard to find prices. Google it. If you think it's going to be remotely reasonable, well, having seen like 5 DRAM launches in prior years. If you've watched prior DRAM launches, you know what's up: they always are overpriced and underperforming at launch. In fact, at launch, every new DRAM technology is priced into the stratosphere at LAUNCH. There has never been an exception. Period. But what will happen is prices will slowly normalize and speeds will increase over the next year or two after launch of the intel platform using it. But, you say, an intel slide told you that prices won't be expensive. Okay. Believe that. They said the same thing about RDRAM. "It wont' have a significant price premium". LOL. Hoooooooooboy did that not happen.

DDR4 will not be an exception. But by all means, keep thinking otherwise. I'll be looking on with amusement when Haswell-E does launch and everyone and their brother is whining about launch DDR 4 prices. You can buy them for servers NOW. 16GB for 600$ of DDR4 at slower rated speeds than comparable DDR3. Have fun with that.

Anyway, Haswell-E will be a great platform for those wanting hexa core or 8 core CPUs. The 5820K should be a great chip. However, make no mistake, DDR4 will be a significant setback for many wanting to go the HEDT route. That will eventually change, but it's going to take at least a year after Haswell-E launches for prices to not be stupid. Most prior DRAM launches took more than a year for prices to go into the sane realm. DDR3 wasn't cheap at launch. Same crap happened. Same crap will happen again.

Then again, does it really freakin matter? Probably not. People adopting the HEDT platform aren't penny pinchers anyway. So in the end it doesn't make a difference. Many guys buying into HEDT area already spending 400$ on a mobo and 600$ on a CPU. Will 500$ for DRAM matter to these guys? Heck no. DDR4 is going to be expensive for the budget minded guys, but most HEDT users simply won't care and will buy it anyway despite the cost.

Before Ram Launch:
Prices Low
Performance better than previous Ram

After Launch:
Prices vastly higher than anyone predicted
Performance worse than previous Ram

But hey, by all means guys, please, pick up DDR4 ram on launch day. Early RAM Adopters help drive down the prices for everyone else so that when the ram actually becomes faster than what was previously out, we can all afford it by then.