• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

[AT] Haswell Refresh comes with improved TIM, unlocked Pentium due mid-2014

Page 11 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
ASRock has been very impressive over the past few years with their motherboards both low and high end. I used to praise Asus, Abit (yes that long ago), MSI and occasionally ECS (laugh if you wish) but I think ASRock is my go-to brand now along with Gigabyte for both Intel and AMD.
 
ASRock has been providing some outstanding values. I still mostly use Giga and Asus, but in a restricted budget situation ASRocks really come in handy.
 
It looks Z97 mobos will be designed specifically for the value minded unlocked Pentium in mind. It's not out yet but ASrock announced they're making anniversary edition Z97 boards which will be value based and cost effective for the pentium specifically. I'd say 80$ range, and other mobo makers will follow suit.

So, as mentioned, there will be value oriented Z97 boards to pair with the unlocked Pentium.
 
It looks Z97 mobos will be designed specifically for the value minded unlocked Pentium in mind. It's not out yet but ASrock announced they're making anniversary edition Z97 boards which will be value based and cost effective for the pentium specifically. I'd say 80$ range, and other mobo makers will follow suit.

So, as mentioned, there will be value oriented Z97 boards to pair with the unlocked Pentium.

cant use a Z87 to o/c the G3258?

cant wait!

http://www.microcenter.com/product/434210/G3258_32_GHz_LGA1150_Boxed_Processor_-_PREORDER
 
Is it me, or is the desktop chip space staring to not be boring as molasses?
It's not just you. The only thing I'm concerned about is whether or not Intel will continue making processors tailored to enthusiasts... you can't just do something big like this and take it away next generation.
 
It's not just you. The only thing I'm concerned about is whether or not Intel will continue making processors tailored to enthusiasts... you can't just do something big like this and take it away next generation.

It'll all depend on sales. If people buy the 4790K/unlocked Pentium in droves, then of course they will continue the momentum.

Honestly, though, I hope that the HEDT platform gets more exciting. I like what I see with Haswell-E, but I want to see follow-through.

SNB-E was disappointing, IVB-E was mostly fail (unless you were moving to it from a 2600K like I was) but HSW-E looks good. I hope with BDW-E and beyond Intel steadily pushes up the core counts.
 
The problem with HEDT is not with the platform itself, which is excellent. The problem is software. Productivity software generally such as video editing and that sort of thing clearly benefits from HEDT. But, since nearly everyone here is of the PC gaming mindset, HEDT isn't a clear cut answer yet. IPC still matters for gaming, and it's clear that 4C is the sweet spot for most games. For every highly threaded game which is 3-4 fps faster on HEDT, you'll find other games that are faster on mainstream because of better IPC.

I do hope software catches up, though, then the answer will be much more clear.

Oh, I do think the 5820k unlocked CPU will spice things up a lot. So I think Haswell-E will be interesting mainly for that: If motherboard prices are reasonable, and once DDR4 prices go way down, then I could see the 5820K being a very hard decision versus the DC. I could see a lot of people buying that chip. The main problem with HSW-E will be DDR4 prices, though.
 
Last edited:
Really sucks that the Pentium will only be overclockable on the Z97, the cheapest motherboard will be a bit more then the Pentium most likely.

Would someone rather spend about $50 maybe on a 1150 mobo with a i3 of their choice or spend a little bit on a Pentium and a bit more on a Z97?
 
Perhaps some will rationalize the need for Z97 by planning a future upgrade to 4690K. I'd guess the Pentium K would have decent resale. It'd be a budget way to get introduced to OCing on these new platforms.
 
It'll all depend on sales. If people buy the 4790K/unlocked Pentium in droves, then of course they will continue the momentum.

Honestly, though, I hope that the HEDT platform gets more exciting. I like what I see with Haswell-E, but I want to see follow-through.

SNB-E was disappointing, IVB-E was mostly fail (unless you were moving to it from a 2600K like I was) but HSW-E looks good. I hope with BDW-E and beyond Intel steadily pushes up the core counts.
Here's an idea... take the packaging tweaks from DC... and apply them to HSW-E :awe:

But yeah, it's definitely unfortunate that the "E" platform has been so neglected. I can see why core counts stagnated though... Intel used massive caches, and Sandy Bridge cores are much beefier than Nehalem. They also moved the system agent/other garbage on die, which took up a tremendous amount of space. Despite that integration... the entry cost was still expensive, and the 4C/8T models were basically useless for most.

Gulftown was 248mm2... SNB-E 435mm2. IVB-E is back down to 257mm2. HSW-E will probably be somewhere in the low 300mm2 region (~325mm2?) assuming a native 8C die.

Take HSW-E, stick it on 14nm, and you'd be looking at maybe 160mm2 with less than perfect scaling. If Broadwell-E is a 10 core die, it'd sit around 180mm2. Intel will likely want to keep the die size around the 200mm2 figure -- 14nm is expensive per mm2. They could potentially integrate the PCH instead, which would probably end up being the same size as a 10C die, or do both and keep it at 200mm2.

I'd imagine that 12 cores is probably asking for too much.
 
Here's an idea... take the packaging tweaks from DC... and apply them to HSW-E :awe:

But yeah, it's definitely unfortunate that the "E" platform has been so neglected. I can see why core counts stagnated though... Intel used massive caches, and Sandy Bridge cores are much beefier than Nehalem. They also moved the system agent/other garbage on die, which took up a tremendous amount of space. Despite that integration... the entry cost was still expensive, and the 4C/8T models were basically useless for most.

Gulftown was 248mm2... SNB-E 435mm2. IVB-E is back down to 257mm2. HSW-E will probably be somewhere in the low 300mm2 region (~325mm2?) assuming a native 8C die.

Take HSW-E, stick it on 14nm, and you'd be looking at maybe 160mm2 with less than perfect scaling. If Broadwell-E is a 10 core die, it'd sit around 180mm2. Intel will likely want to keep the die size around the 200mm2 figure -- 14nm is expensive per mm2. They could potentially integrate the PCH instead, which would probably end up being the same size as a 10C die, or do both and keep it at 200mm2.

I'd imagine that 12 cores is probably asking for too much.

Great post, Homeles.
 
The problem with HEDT is not with the platform itself, which is excellent. The problem is software. Productivity software generally such as video editing and that sort of thing clearly benefits from HEDT. But, since nearly everyone here is of the PC gaming mindset, HEDT isn't a clear cut answer yet. IPC still matters for gaming, and it's clear that 4C is the sweet spot for most games. For every highly threaded game which is 3-4 fps faster on HEDT, you'll find other games that are faster on mainstream because of better IPC.

I do hope software catches up, though, then the answer will be much more clear.

Oh, I do think the 5820k unlocked CPU will spice things up a lot. So I think Haswell-E will be interesting mainly for that: If motherboard prices are reasonable, and once DDR4 prices go way down, then I could see the 5820K being a very hard decision versus the DC. I could see a lot of people buying that chip. The main problem with HSW-E will be DDR4 prices, though.

By the time DDR4 pricing is on par with DDR3 pricing, the next HEDT platform will be out. For me it's what comes after Haswell-E that's interesting. However, with a $300 processor promising OCs of 5Ghz, I don't see much need for "MORE CORES!!!!!!!!" when I'd gain far more performance from DC.

There just aren't enough applications where more cores really benefits me.

The majority of people even begging for more cores don't really have practical EVERYDAY situations that they need more cores for. It's just typical enthusiasts that want more of everything just for the sake of having it.
 
Really sucks that the Pentium will only be overclockable on the Z97, the cheapest motherboard will be a bit more then the Pentium most likely.

If you look at post #243, the bottom slide does say Pentium 20th Anniversary is compatible with 8 series.

So it should be overclockable on Z87. (Check for BIOS support though)
 
Here's an idea... take the packaging tweaks from DC... and apply them to HSW-E :awe:
IVB-E already had superior packaging; it was designed for greater power delivery and the heatsink was soldered. I don't see why HSW-E would be any different.
 
Regarding "improved TIM" and "die to heatspreader distance" problem......why not BGA packaging for future unlocked dual core overclocking processors?

(A dual core, even if overclocked, shouldn't need much of a cooler. So even a relatively small (by aftermarket tower cooler standards) BGA cooler should work fine I'd imagine)
 
Last edited:
http://pclab.pl/art57691.html

pentium_g3258_4700.png


fsx_1920n.png


wd_1920n.png


as expected, great for some things, very bad for others...
 
Except for very special use cases, I don't really see the appeal of the Pentium. Just not well rounded. Now a 4.6 ghz i3 would be another story, but that would have eaten too much into i5 sales.
 
I'm going to push for 5ghz with the Anniversary Pentium. This endeavor is going to provide me some fun, that's all. A pleasant memory, hopefully!
 
Back
Top