Leaked Intel Haswell Processor Refresh Prices?

HardOCP News

[H] News
Joined
Dec 31, 1969
Messages
0
The folks at CPU-World have the lowdown on what they say are prices for Intel's Haswell processor refresh. Not a whole lot to get excited about but that 4GHz i7-4790 looks tasty.
 
So this is just a 100Mhz speed bump so Intel can tell shareholders they're "continually rolling out new products"?

Wow. Fun.
 
Maybe...just maybe...Intel realized that they should have been soldering the heatspreader to the die the entire time, or at the very least doing it for the consumer labeled K parts that aren't -E or Xeon. One can dream.
 
Looking at the prices it tells me the 4Ghz is actually the max clock (Turbo) and not the base clock... so yeah... big meh :(
 
Yawn.. I guess we will wait until later this year for some real new stuff to come out.
 
Another year without a wish to upgrade.
C'mon Intel.
 
Mine as well wait for X99 and see what comes after that. Saving a bit more money this year so far!
 
Not sure why everyone is surprised/disappointed. Maybe I missed the news and crazy rumors that haswell refresh was going to be revolutionary upgrade? I can understand the disappointed in Haswell since it was hyped up a bit, but haswell refresh?
 
Not sure why everyone is surprised/disappointed. Maybe I missed the news and crazy rumors that haswell refresh was going to be revolutionary upgrade? I can understand the disappointed in Haswell since it was hyped up a bit, but haswell refresh?

I think many of us, even though we've heard it mentioned, just can't understand why Intel would think this is a useful thing to do. It's a complete waste. These products are nowhere near enough of an improvement for them to bother, and a lot of us get annoyed when Nvidia makes similar moves (releases a new product to try and get people to upgrade, but the new product isn't actually a real upgrade). It makes sense for Intel to slow down the releases, though, considering the lack of competition and the rising costs of R&D as process nodes shrink.

Apparently OEMs must think that their new machines need to keep delivering new CPUs to sell (or Intel feels people have gotten too used to it's not-really-yearly processor releases), when what they really need is to stop using Windows 8.

They can keep this one. I'm not buying, and I'm not recommending anyone upgrade to it (unless their current machine is so bad that they can't wait for Broadwell).
 
Any new featureset in these processors? Features > clockspeed these days.
 
I think many of us, even though we've heard it mentioned, just can't understand why Intel would think this is a useful thing to do. It's a complete waste. These products are nowhere near enough of an improvement for them to bother, and a lot of us get annoyed when Nvidia makes similar moves (releases a new product to try and get people to upgrade, but the new product isn't actually a real upgrade). It makes sense for Intel to slow down the releases, though, considering the lack of competition and the rising costs of R&D as process nodes shrink.

Apparently OEMs must think that their new machines need to keep delivering new CPUs to sell (or Intel feels people have gotten too used to it's not-really-yearly processor releases), when what they really need is to stop using Windows 8.

They can keep this one. I'm not buying, and I'm not recommending anyone upgrade to it (unless their current machine is so bad that they can't wait for Broadwell).

It's pretty much the same thing Intel did with the 2700k... In other words, nothing to really see here.
 
Still rocking my Sandybridge 2500k @4.5GHz. Don't see any compelling reason to upgrade on the horizon.
 
Well, since my fastest CPU is an Intel E8500, I find the 4590 and 4690 pretty interesting. It's nice to see what others are thinking though. I didn't know Intel progress was slower lately.
 
Honestly I wasn't expecting anything more for Haswell, so even a little bump is nice I guess.
 
Maybe...just maybe...Intel realized that they should have been soldering the heatspreader to the die the entire time, or at the very least doing it for the consumer labeled K parts that aren't -E or Xeon. One can dream.
In either case, worse TIM or dealing with a higher power density in contact with IHS, Intel needs to find another solution in the next node or two. It might not necessarily mean going back to solder, but may require something better.
 
People seem a bit confused about what these refreshes are.

The clock speed for a CPU is based heavily on the stability they see from sample batches of silicon.
A smaller sample size in production gives a larger amount of uncertainty with stability, so they label them lower off the lot.

Now, having a larger sample size, their uncertainty is lower and they can comfortably sell the silicon with higher clock ratings.
 
People seem a bit confused about what these refreshes are.

The clock speed for a CPU is based heavily on the stability they see from sample batches of silicon.
A smaller sample size in production gives a larger amount of uncertainty with stability, so they label them lower off the lot.

Now, having a larger sample size, their uncertainty is lower and they can comfortably sell the silicon with higher clock ratings.

Further, what is exactly pushing existing (or even dead) hardware?

I run three operating systems in rotation - Windows 8.1 ProWMC (x64), Windows Server 2012R2 (also x64) and (currently) openSuSE 13.1 (x64) - and I'm doing it on dead hardware. (openSuSE has replaced, for now, OS X Mavericks 10.9.2.)

Other than niche software, there's not exactly anything really pushing my current hardware - heck, even Titanfall beta didn't. Worse, whenever someone even tries to push the envelope, there has been massive pushback. (Crytek, Microsoft, etc.) As much as some enthusiasts would wish that Intel/AMD/nVidia would push the hardware envelope, why should they when the vast consumer majority could care less?

I'm still going to upgrade - however, it's driven by outlier usage, not anything mainstream. Worse, I know it, and I refuse to blow smoke up anyone's posterior and claim otherwise when I know better.

In order to push the envelope, there has to be a reason for it - what enthusiasts want is not enough.
 
What you all seem to be missing the flip chip description code...

LGA12C!!!

They're all 12 core CPU's!!! z0mg!!!

















;)
 
People seem a bit confused about what these refreshes are.

The clock speed for a CPU is based heavily on the stability they see from sample batches of silicon.
A smaller sample size in production gives a larger amount of uncertainty with stability, so they label them lower off the lot.

Now, having a larger sample size, their uncertainty is lower and they can comfortably sell the silicon with higher clock ratings.

What you said can be summed up by saying that Intel's yields have improved, more specifically yields of the top bins has improved and this yield improvement permits higher speed CPUs. These new SKUs (and the new chipsets) are Intel's way of saying, "We're still moving forward" when the products are in actuality nearly identical to their predecessors.
 
I hope for the OC/Enthusiast community's sake they do something shocking like solder the IHS. It's sad that we're at the point where that actually would be "shocking".

The realist in me however tells me it's just going to be a re-brand with a 100MHz boost. I hope I'm wrong but we'll see!
 
I hope for the OC/Enthusiast community's sake they do something shocking like solder the IHS. It's sad that we're at the point where that actually would be "shocking".

The realist in me however tells me it's just going to be a re-brand with a 100MHz boost. I hope I'm wrong but we'll see!

It would be nice if they can soldered by default, but that would be a few extra cents per CPU.:rolleyes:

You can order the thermal pads and do it yourself...Honeywell PCM45F

http://www.ebay.com/itm/Honeywell-PCM45F-Heatsink-Compound-in-sheet-form-10-pcs-/390351101901

Cheap and shouldn't be too hard to accomplish. Phase change at 45°C...should be fairly easy for a regular heat gun to generate with no risk to the chip. Get a few strong clothespins, clamp the thermal material between the IHS and the die and let 'er rip...;)
 
As an eBay Associate, HardForum may earn from qualifying purchases.
I hope for the OC/Enthusiast community's sake they do something shocking like solder the IHS. It's sad that we're at the point where that actually would be "shocking".
The situation is especially appalling in the "K" versions of their recent chips.
 
The situation is especially appalling in the "K" versions of their recent chips.

As all LGA1150 chips likely come off the same assembly line, both the locked and unlocked versions, I doubt Intel has the ability to solder only a select few CPUs that come down that line. It's either all of them or none of them. That said, the cost to solder them is really only pennies per CPU. Intel is just being ridiculously (and inexcusably) cheap here by not soldering them all.
 
Other than niche software, there's not exactly anything really pushing my current hardware - heck, even Titanfall beta didn't. Worse, whenever someone even tries to push the envelope, there has been massive pushback. (Crytek, Microsoft, etc.) As much as some enthusiasts would wish that Intel/AMD/nVidia would push the hardware envelope, why should they when the vast consumer majority could care less?

Anyone streaming to Twitch wants better CPUs, both speed and cores. And Twitch isn't exactly "niche" any more. x264 encoding can utilize over 20 cores before diminishing returns kick in, and that's at 1080p, 4K will be able to use over 40 cores before diminishing returns. i7s are significantly better for live encoding while gaming than the i5s that gamers seem to pick for value reasons. The market for this is already large and growing more than most other PC markets. On a somewhat related note, PCs are remarkably better at this than laptops, let alone handhelds (which are wholly unable). Streaming software like Open Broadcaster Software is more than able to push any CPU to its limit, not even the best Xeon on the planet can satisfy the x264 encoder at 1080p; there's always a higher quality preset.
 
Solder myth is still going around ?

I was thinking it would be dead after we saw Ivy-E not overclocking any better than 1155 Ivy
 
Solder myth is still going around ?

I was thinking it would be dead after we saw Ivy-E not overclocking any better than 1155 Ivy
I agree, but it's possible to overclock to marginally higher clock speeds (without hitting throttle limits) by replacing the TIM Intel is using. Intel could do better than what it ships in K models at least.
 
Solder myth is still going around ?

I was thinking it would be dead after we saw Ivy-E not overclocking any better than 1155 Ivy

My memory is a bit hazy on the subject, but I thought it did OC slightly better, with lower temps but still couldn't match SB's overclocking potential?
 
Anyone streaming to Twitch wants better CPUs, both speed and cores. And Twitch isn't exactly "niche" any more. x264 encoding can utilize over 20 cores before diminishing returns kick in, and that's at 1080p, 4K will be able to use over 40 cores before diminishing returns. i7s are significantly better for live encoding while gaming than the i5s that gamers seem to pick for value reasons. The market for this is already large and growing more than most other PC markets. On a somewhat related note, PCs are remarkably better at this than laptops, let alone handhelds (which are wholly unable). Streaming software like Open Broadcaster Software is more than able to push any CPU to its limit, not even the best Xeon on the planet can satisfy the x264 encoder at 1080p; there's always a higher quality preset.

This. I can't justify upgrading my gulftown encoding box right now as I'm not sure if I'd even get any noticable gains. The processing power required to go up a noticable amount is huge and incremental over the last setting - meaning we just need exponentially more power than we have available if we want to broadcast higher quality video than we can now at reasonable bitrates. I can get it close to realistic in most games, but in higher-movement titles the pixelation is still present.

x265 should be out this year or next, and that will help me get more quality out of my stream. So there's really no reason to "gamble" on an upgrade right now that may or may not do anything for me, especially when software upgrades are always coming for constant improvements.

Eventual pressure from AMD is also a factor. I'm impatient, but not impatient enough to buy those price-gougey Xeons. Inflation is a cool excuse and all, but it's not the reason a noticable high-end computer upgrade costs what a car costs. That's greed - any company would do it if there was no competition.

If there was a decent 8c/16t or more avalable at a reasonable price(500-1k, typical of high-end intel consumer silicon), I'd order parts for a new box tonight.
 
Anyone streaming to Twitch wants better CPUs, both speed and cores. And Twitch isn't exactly "niche" any more. x264 encoding can utilize over 20 cores before diminishing returns kick in, and that's at 1080p, 4K will be able to use over 40 cores before diminishing returns.

we just need exponentially more power than we have available if we want to broadcast higher quality video than we can now at reasonable bitrates

I am sorry (not to mention rather ignorant) but I am not sure I follow this whole "need more CPU power to stream" business.

If one were that serious about streaming then why not invest in an NVIDIA GPU and call it a day? NVIDIA's NVREC/ShadowPlay have such a minimal impact on overall performance that it seems like a no brainier if one were into streaming. If anything, what WOULD cause your streams to come out blocky and ugly is not so much GPU vs. CPU encoding but rather the 3500kb/s limit imposed by Twitch. One could argue that if you were that far into streaming that you would have already invested in an external PVR, no?
 
My 2600K is still running strong, and I wouldn't gain a whole lot by upgrading. Intel used to have me biting at the bit for a new CPU every couple years. It's been a few, and I'm still not excited about the new ones. Guess I'll wait another year to upgrade. At least I'm definitely getting my moneys worth on this build. It's lasted quite a while, and still plows through anything I throw at it.

Sucks, too. I like upgrading. Intel could have had my cash if they had something exciting to show.
 
My 2600K is still running strong, and I wouldn't gain a whole lot by upgrading. Intel used to have me biting at the bit for a new CPU every couple years. It's been a few, and I'm still not excited about the new ones. Guess I'll wait another year to upgrade. At least I'm definitely getting my moneys worth on this build. It's lasted quite a while, and still plows through anything I throw at it.

Sucks, too. I like upgrading. Intel could have had my cash if they had something exciting to show.
Well said, Sir.

Sitting pretty here myself with a 2500K build @4.5Ghz. At this point, the upgrade bug is at full blown infestation level. However, I'll wait another generation before upgrading. Ivy and Has = meh.
 
What do you want to see in a new mobo?

I'm not really sure--it seems silly I know but I don't like anything out there with bluetooth... I think I'd like a new manufacturer to suddenly appear but I guess I wouldn't mind trying out a nice EVGA 1150 board. The only boards I'd be interested in using for builds instead of the Hero would be the Asrock boards and they're just not laid out to my liking and as an alternative to Asus, Asrock doesn't really seem like it's quite separate enough from their old daddy. Still feels like buying from Asus.
 
I'm waiting for Haswell-E with X99. It'll be a great upgrade from socket 2011 systems.
 
Back
Top