IVY-E!!

A

Arkanian

[H]ard|Gawd
Joined
Feb 8, 2004
Messages
1,898
So looking forward to this chip. With the recent leak of it using solder instead of TIM I am hoping for some 5 GHz over clocks! Sept can't get here fast enough!
 
Looking forward to some high temps eh? Lol 5 ghz not gonna happen unless you find an extremely golden one.
 
As long as they stick to the solder for the retail chips I could see quite a few chips being capable of it. It is the tim that ruined temps on ivy and haswell.
 
spaceace420 said:
As long as they stick to the solder for the retail chips I could see quite a few chips being capable of it. It is the tim that ruined temps on ivy and haswell.
Click to expand...

I think you're going to be disappointed. There's plenty of delidded IBs and Haswells out there and very few are capable of 5GHz on every day cooling.
 
Can't be any hotter than the 9xx series. I'm still using one at work with the giant stock Intel heatsink/fan on a 970 and that pushes some warm air!
 
spaceace420 said:
As long as they stick to the solder for the retail chips I could see quite a few chips being capable of it. It is the tim that ruined temps on ivy and haswell.
Click to expand...

Agreed. Both of my 3770ks can do 5ghz but they hit TJ Max under load. They would be 24/7 5ghz if they had solder.

However only 6 cores still is making me hesitate on upgrading. Still rocking a i7-970 for my encoding box and still not seeing a compelling reason to upgrade to a system with the same amount of cores.

Something about buying "new" tech that is actually year old architecture really rubs me the wrong way. I'm not interested in sacrificing anything when I'm buying a $1k+ processor, and right now intel is making us sacrifice IPC vs Haswell, and not all of that will be made up for with the increased OC headroom.

Edit: Also no AVX2, which has started to get application support and it does indeed make a huge difference in encoding speed. A $300 Haswell chip has it but IB-E chip will not. I understand the logic as to why they did this, but understanding it isn't making me want to buy IB-E as a consumer even though I am the target audience for this product..
 
Last edited:
I'm sorry is there something wrong with my 2600k? 5 less watts and 5% more IPC in exchange for lower clocks? Dang I'm missing out.
 
Intel is just one year too late to put Ivy Bridge Extreme on 22nm, should have done it last year and not to mention its most likely going to be only 10% faster or less than the top end LGA 2011 offerings.

And if they use that cheap TIM just like Ivy Bridge and Haswell... epic fail.
 
Parja said:
I think you're going to be disappointed. There's plenty of delidded IBs and Haswells out there and very few are capable of 5GHz on every day cooling.
Click to expand...
Solder is better than any TIM on the planet.
 
Tsumi said:
Direct die cooling should theoretically be better than solder and an IHS.
Click to expand...

Unless you use a metal based TIM like the CL Ultra like I do, then a soldered IHS would be better since you have a great surface area for the weaker TIM to transfer the heat from..

Personally I wouldn't run any modern cpu that doesn't have a soldered IHS with it on..I was beyond happy to remove mine and mount my cpu block directly to the die with CL Ultra..
 
ccityinstaller said:
Unless you use a metal based TIM like the CL Ultra like I do, then a soldered IHS would be better since you have a great surface area for the weaker TIM to transfer the heat from..

Personally I wouldn't run any modern cpu that doesn't have a soldered IHS with it on..I was beyond happy to remove mine and mount my cpu block directly to the die with CL Ultra..
Click to expand...

Youre still limited by the surface area of the die. The less between the die and cooling solution, the better. Solder and an aluminum IHS have worse thermal conductivity than copper.

Die > IHS > TIM > copper heatsink

will always be inferior to...

DIE > TIM > copper heatsink
 
I can't wait for IB-E and glad they're using solder. Although maybe too hot for ambient cooling to really oc like crazy but should be great for sub-zero.
 
I'm really not looking forward to IB-E. Unless of course there will be an 8-core one. Then I will be super happy.
 
Hoping Intel brings some IB-E goodness that convinces me to upgrade. Still feeling like my 930 is holding up just fine against current offerings.
 
SpeedyVV said:
I'm really not looking forward to IB-E. Unless of course there will be an 8-core one. Then I will be super happy.
Click to expand...

The cheapest (2 GHz) 8-core Xeons are ~$1100, I wonder what they would charge for an unlocked 8-core...

It doesn't really seem worthwhile to spend a whole lot of money for two more cores unless you encode video all day.
 
I'll definitely give IB-E a go. Don't need the upgrade for gaming but am really psyched to see how they do cold.
 
I have so much water cooling capacity its overkill and rediculous.

I can hit 5ghz on my sandy 3930K. Stop pissing your lack of knowledge on this dudes thread.

At this point fuck 6 or 8 or 10 cores...

Give me 1 core that does 150ghz and now were talking POWER!
 
tangoseal said:
I have so much water cooling capacity its overkill and rediculous.

I can hit 5ghz on my sandy 3930K. Stop pissing your lack of knowledge on this dudes thread.

At this point fuck 6 or 8 or 10 cores...

Give me 1 core that does 150ghz and now were talking POWER!
Click to expand...

U drunk?
 
I'll pass untill they release Haswell-E since IB-E will still use those old X79 chipsets.
 
Seriously, IB-E silicon has 12 cores, yet we still only get 6-core unlocked i7s. I was pissed at intel disabling two cores on SB-E but I could ALMOST understand it, but now a $1000 i7 is literally a piece of junk silicon with SIX disabled cores. At least give us 8 core i7 chips!
 
Wont even consider upgrading my 3930k unless more cores are available... what is Intel thinking?

At least give me an 8 core chip around $500-600 and you've got my money... anything else is a waste .
 
Tsumi said:
Direct die cooling should theoretically be better than solder and an IHS.
Click to expand...

nissanztt90 said:
Youre still limited by the surface area of the die. The less between the die and cooling solution, the better. Solder and an aluminum IHS have worse thermal conductivity than copper.

Die > IHS > TIM > copper heatsink

will always be inferior to...

DIE > TIM > copper heatsink
Click to expand...

[H] solder their high end waterblock directly to the die.

(I actually plan to do this. :-P )
 
flegg said:
I'm sorry is there something wrong with my 2600k? 5 less watts and 5% more IPC in exchange for lower clocks? Dang I'm missing out.
Click to expand...

That's where I'm at. 2600k that runs just fine at 4.5 GHz. I like the new chips, and they are awesome. But, they just aren't as much of a leap as previous gen.

You might hit 5 GHz, but it wouldn't be too common. I'd love to see it be a common thing (and I might upgrade if it's a good success rate). But, unless it's a great chip with great batches, I'll pass.
 
Isn't DDR4 coming out with the Haswell refresh/Z97 which is supposedly Q2 '14? Why get an Ivy-E now when a few months later you can get a refresh and DDR4?
 
King Icewind said:
Isn't DDR4 coming out with the Haswell refresh/Z97 which is supposedly Q2 '14? Why get an Ivy-E now when a few months later you can get a refresh and DDR4?
Click to expand...

Unless things have changed, it'll still be a 4-core, and an incoming version of DDR generally has piss poor timings compared to the outgoing version running at the same speed (e.g. DDR2-1066 with 5-5-5-15 vs DDR3-1066 with 7-7-7-21).
 
King Icewind said:
Isn't DDR4 coming out with the Haswell refresh/Z97 which is supposedly Q2 '14? Why get an Ivy-E now when a few months later you can get a refresh and DDR4?
Click to expand...

This will use DDR3.


Haswell-E on the new pin incompatible LGA2011 will use DDR4 but that will come out in Q3 or Q4.
 
ryan_975 said:
Unless things have changed, it'll still be a 4-core, and an incoming version of DDR generally has piss poor timings compared to the outgoing version running at the same speed (e.g. DDR2-1066 with 5-5-5-15 vs DDR3-1066 with 7-7-7-21).
Click to expand...

No official word of cores to my knowledge, but since it's a HWL revision and not BWL, I will side with you on it being 4 core max only.

As for DDR4, I really don't think memory speed and timings are going to make much of a real-world difference with the upcoming platform, just like they don't with recent/current platforms based on Patsburg, Cougar Point, Panther Point, and Lynx Point.
 
You must log in or register to reply here.
Back
Top