Sign in with
Sign up | Sign in
Your question
Closed

Ivy Bridge or... Haswell?

Tags:
Last response: in CPUs
Share
April 24, 2012 4:47:36 AM

Let me start by saying that I don't upgrade my PC often, as is evident by my current 5-year-old rig featuring a Q6600@3.6. Modern games are playable, but not always enjoyable. As a result, the itch towards a new build is ever-strong.

In the past couple of weeks I've bought a tower (Fractal Define XL) and PSU (Corsair HX850) in anticipation for a new Ivy Bridge build, along with major fan upgrades and a fan controller. I'm around 450 USD deep into the build.

I didn't take the plunge with Sandy Bridge (not an option for me at this stage), but I was impressed with the gains over the previous architecture. Although I understand it represents a different stage in Intel's lineup, I was expecting something equally significant with Ivy Bridge. I'm admittedly a little disappointed with the less-than-stellar initial reviews of Ivy Bridge, particularly the relatively low overclock ceiling. While a vast improvement over my current machine is certain, I'm now finding it hard to justify the upgrade, when it seems like the improvement is minimal compared to last year's Sandy Bridge and its predecessor.

So, considering I've started the build already (granted with components that depreciate the least), do I wrap up this new case/PSU and throw it in the closet for a year to see what Haswell has to offer, or stick with "Plan A" and buy the Ivy Bridge architecture despite my reluctance?

My brain is going in circles right now and I'm just looking for a little outside perspective. Please advise.

More about : ivy bridge haswell

a b à CPUs
April 24, 2012 4:48:52 AM

If you can wait, I would go for Haswell. Intel should have improved the relatively new 22nm process by then, and should be a lot better than Ivy.
Score
0
a c 186 à CPUs
April 24, 2012 4:53:41 AM

Core 2 Quad to haswell is the best way. If possible, up your cooling and aim for 4ghz.
Score
0
Related resources
April 24, 2012 5:19:09 AM

If I were you I would wait for H-as-well-.
Score
0
April 24, 2012 6:08:32 AM

Why wait for Haswell? I'm getting IB and i've been running my PC for 5 + years. By the time I need a new PC with IB haswell will be gone and it will be on to the next big architecture.
Score
0
a b à CPUs
April 24, 2012 6:40:42 AM

Try ride out another year or so.
Score
0
April 24, 2012 7:34:03 AM

Curious why everybody said to wait?
Some years ago, when I jumped from q6600 to q9650, it was very nice bump in performance... the SB/IB are so much better than my q9650 that to me is a no brainer to go SB/IB if one can/want...
Score
0
a c 478 à CPUs
April 24, 2012 7:34:27 AM

Waiting for Haswell. My Q9450 should be about 5 years old by then...
Score
0
April 24, 2012 9:05:04 AM

There will always be the "next" thing to come around...
When you decide you need a new computer, buy a new computer.
One of the worst things to do is buy parts like 6-9 months in advance because by the time you assemble it together there's 1: something newer and better, 2: have wasted so many months of the warranty...

Just my 0.02C
Score
0
a b à CPUs
April 24, 2012 9:17:20 AM

The downside to going to ivy is that you will be on a dead socket after a year, it is true the chip will be good for the next 4-5 years, but it is always the decision to make when moving up.

I do concur that a jump to IB is significant.
Score
0
a c 80 à CPUs
April 24, 2012 10:26:09 AM

You'll always be on a dead socket within a 2 year timeframe, 775 (in name only) lasted longer, but 1156, 2011, 155 all going to be dead, along with whatever haswell's socket will be, that'll declared dead by 2014/15 and probably replaced by 2015/6. AMD were no better with the AM sockets being compatible, but to get the high performing chips you needed to be on the latest iteration, FM1 dead.

They may have to do it to keep mobo sales up, and it wouldn't suprise me to see an unconventional socket for haswell so that the heat can be removed quicker, if a HS could be attached to both sides of the ship you are doubling the surface area.

If you want the performance bump, and can afford it, and want to afford it, just do it.
Score
0
!