More Leaked Benchmarks from Intel's Ivy Bridge

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.
[citation][nom]maxinexus[/nom]Who cares about liquid N2 performance..it something for 0.00000000001% of people and the rest of the planet simply wants the chips up for sale already[/citation]

It proves that when taken to the extreme Ivy Bridge has the chops to clock that high and perform well still.
 
And Nvidia's saying "no no, you exaggerate the scale and make it look like the increases are eleventy bajillion percent. Silly intel!"
 
[citation][nom]southernshark[/nom]Yeah I'm still gaming fine with my Phenom II x 6. I'll upgrade either when I need it for gaming or when the CPU companies come out with something truly impressive. 10-15 percent gains don't make me want to upgrade (especially when I won't notice much if any performance in games). Even if games don't improve though I would upgrade the CPU if it really impressed me. Right now though I don't see the point of upgrading.[/citation]

Ivy has about 50% more IPC than Phenom II does and it clocks much higher. Obviously not a 10-15% gain there. Even Sandy has about 40% more IPC than Phenom II and clocks higher, so it's still not a 10-15% gain there either. You can even go back as far as Nehalem and it still has like 20-30% more IPC than Phenom II and can clock at least as high, if not higher (depending on the CPU), so even Nehalem is more than 10-15% faster than Phenom II. I have a Phenom II x6 1090T BE system (not a gaming system) and I can confirm all of this myself.

[citation][nom]nao1120[/nom]What hes trying to say is and i agree with - There is no reason to upgrade. If you have a fast Quad, Either the Q9550(currently have from 2008, Technology from 2007 ish), AM3 - T1090, or T1100 phenom 2, or a first gen i7 860-920 there is no point.Anything as a base quad core at 2.8GHZ- 3.2 will be fine. I think it'll still be fine going into 2013. And no.... I don't overclock. Use a budget 6870. Feel free to spend all your cash![/citation]

This all depends on exactly what you do and how you do it with the computer. If I have a dual 7970 setup, then even a highly overclocked 2500K may still bottleneck performance a little in most games. Same goes for a dual GTX 680 setup. There's even a noticeable difference between the i5-2500K at stock and overclocked to 4GHz in some games with a single 7970 or GTX 680 and even with cheaper graphics cards.

If I had a Core 2 quad instead of a Phenom II x6, then I would easily notice a difference in performance for non-gaming work because the Phenom II x6 has 50% more cores and each core is about as fast as most Core 2 cores are (little less IPC, but not by too much). I'd say that at 4GHz, my 1090T BE is about 35-40% faster than a Core 2 Quad 9xxx at 4GHz.

For mid-range gaming like what you do, even the FX-4100 and Nehalem i3s and i5s should be able to handle a Radeon 6870. However, a GTX 80 and i5-2500K@4.x GHz will look FAR better during play than anything with only a 6870 can muster. That's the whole point of high end gaming rigs... They have such greater picture quality (unless paired with a crap monitor or crap monitors). Grab a GTX 680 and playing at 2560x1600 with some AA and maxed out quality settings on a high quality monitor looks simply awesome. A 6870 can't even do 1080p in some of the newer games.

[citation][nom]maxinexus[/nom]Who cares about liquid N2 performance..it something for 0.00000000001% of people and the rest of the planet simply wants the chips up for sale already[/citation]

It shows that the Ivy Bridge CPUs overclock a lot farther than the Sandy Bridge CPUs do. You can expect similar (possibly greater due to the exponential nature of overclocking power usage/heat generation) differences between Sandy and Ivy even with weaker (aka cheaper) cooling setups.
 
[citation][nom]andboomer[/nom]And Nvidia's saying "no no, you exaggerate the scale and make it look like the increases are eleventy bajillion percent. Silly intel!"[/citation]

The scale is perfect on these graphs so that excuse is invalid. There is not cutoff at the start nor end to exaggerate the scale and the graph goes up linearly too so it's not confusing readers as it may were it logarithmic or exponential to decrease/increase the apparent differences shown by the graphs.
 
Intel plx release Ivy Bridge! Im dying here from lack of playing.
My gaming pc from 2008 died on me "the motherboard", and im not going to repair it... Waste of money on a dead socket, gonna buy one ivy bridge core i7 and hd7870 unless nvidia releases something less expensive.
 
[citation][nom]blazorthon[/nom]The scale is perfect on these graphs so that excuse is invalid. There is not cutoff at the start nor end to exaggerate the scale and the graph goes up linearly too so it's not confusing readers as it may were it logarithmic or exponential to decrease/increase the apparent differences shown by the graphs.[/citation]uh, no sh*t Sherlock. That was my point (plus a dig at Nvidia)
 
[citation][nom]doive1231[/nom]Ivy Bridge allows PCI Express 3.0. Soon GPUs could need this bandwidth so a valid reason to upgrade.[/citation]

I guess GPU's "COULD" need PCI Express 3.0 in the near future but the majority will likely not need it. The only people that might see some benefit from 3.0 are people running 2 or 3 top end GPU's, the average person will see no benefit from 3.0 for a long time
 
no news here, I want to see it against the 3960x =)

And yes it can go higher, I bet we will see something @8Ghz on those when the crazy LN2 folks get their hands on them ;p
 
[citation][nom]soldier37[/nom]So bottom line is it worth it to go from my 4.5 ghz 2600k to a 3770 clocked the same, what difference will I see in converting movies and gaming real world percentage 10-15%? Is that really worth $350 upgrade for Ivy?[/citation]
No, and judging by the power use differences, you wouldn't make up the price in power savings. If you are already on SB there is no need for IB unless you like being up to date with the best there is to offer every year. Look at the benchmarks in the link i proved above.
 
[citation][nom]soldier37[/nom]So bottom line is it worth it to go from my 4.5 ghz 2600k to a 3770 clocked the same, what difference will I see in converting movies and gaming real world percentage 10-15%? Is that really worth $350 upgrade for Ivy?[/citation]

3770 can overclock higher with the same cooling because of it's reduced power consumption so it's more like comapring a 4.5GHz 2600K to a 5GHz 3770 when the 3770 has a 10-15% advantage in IPC in addition to the clock advantage. Still, not worth a $350 upgrade (at least not in my opinion), but only because you already bought the 2600K. Had you been looking to build a new computer and this 3770K was an option for $350 and the 2600K is for $300, then the 3770K would be the obvious choice if you're willing to spend more money to get more.

As for the encoding and such, Ivy might be FAR faster than Sandy because of it's greatly improved IGP and might be worth it. However, for CPU performance in productivity and games, I don't think it's worth the money if you already have an i7.
 
[citation][nom]blazorthon[/nom]3770 can overclock higher with the same cooling because of it's reduced power consumption so it's more like comapring a 4.5GHz 2600K to a 5GHz 3770 when the 3770 has a 10-15% advantage in IPC in addition to the clock advantage. Still, not worth a $350 upgrade (at least not in my opinion), but only because you already bought the 2600K. Had you been looking to build a new computer and this 3770K was an option for $350 and the 2600K is for $300, then the 3770K would be the obvious choice if you're willing to spend more money to get more.As for the encoding and such, Ivy might be FAR faster than Sandy because of it's greatly improved IGP and might be worth it. However, for CPU performance in productivity and games, I don't think it's worth the money if you already have an i7.[/citation]

Your guess on OC is pretty safe assumption, if on LN 17-2600k gets 6ghz and the i7-3770k gets 6.9ghz it's roughly a 15% improvement in OC (4.5ghz +15% is 5.1ghz roughly). If this translates well to air cooling, we could see 5ghz be the norm on IB compared to 4.5ghz being the norm on SB.
 
[citation][nom]eddieroolz[/nom]It proves that when taken to the extreme Ivy Bridge has the chops to clock that high and perform well still.[/citation]

No, it doesn't. It proves pretty much nothing, and is a totally meaningless demonstration. Pretty much the equivalent of strapping a rocket to a Volkswagen Beetle. Then concluding, see how fast the Beetle can go? This proves the true potential of Beetle. No, it doesn't. It's totally unusable like that, and their is no correlation between how either product performs under such extreme circumstances and how the average sample will perform for any of us under normal operating conditions.
 
[citation][nom]A Bad Day[/nom]But, CLOCK RATE IS THE ONLY THING THAT MATTERS! The Pentium 4 is faster because it can be OCed to 8 GHz!/sarcasm, do not flame me please[/citation]

Wow! are you stupid! just kidding :) I still have two Pentium 4's laying around... keep them to remember the horrendous failure they were/are. Have one set as a dedicated ddr machine as that's all it's worth :lol:
 
[citation][nom]Kyuuketsuki[/nom]I'm willing to bet the majority of the improvement in those benchmarks come from the higher turbo frequencies.[/citation]

You are talking about the official slide results right? You will see your gains, but only when both of your setups are using integrated graphics. It seems that Sandy Bridge loses few % by using integrated graphics, and Ivy Bridge could fix that. So the results are true, just not using integrated graphics.

It's unlikely that Turbo frequencies will account for more than 1-2% gains in average because desktop CPUs don't give big gains using Turbo Mode.
 
I want IB now 🙁 I want to upgrade my AMD 965 to a 3770K :)

I wonder if my Zalman CNPS9900 120mm copper cooler will cool a 3770K even the IB runs cooler than AMD?
 
[citation][nom]kinggremlin[/nom]No, it doesn't. It proves pretty much nothing, and is a totally meaningless demonstration. Pretty much the equivalent of strapping a rocket to a Volkswagen Beetle. Then concluding, see how fast the Beetle can go? This proves the true potential of Beetle. No, it doesn't. It's totally unusable like that, and their is no correlation between how either product performs under such extreme circumstances and how the average sample will perform for any of us under normal operating conditions.[/citation]

If the i7-3770K can run benchmarks at that high of an overclock, then it obviously isn't useless. Also, it does tell us the ttrue potential. It tells us the absolute best that the processor is capable of. It also tells us that it's best is better than it's predecessors. We can expect a similar benefit in overclocking on any other cooling system such as air and water, if not a greater benefit over Sandy. It also tells us that the 3770K can overclock more stably at the same frequency as Sandy because it can overclock a lot farther than Sandy.

[citation][nom]omega21xx[/nom]Your guess on OC is pretty safe assumption, if on LN 17-2600k gets 6ghz and the i7-3770k gets 6.9ghz it's roughly a 15% improvement in OC (4.5ghz +15% is 5.1ghz roughly). If this translates well to air cooling, we could see 5ghz be the norm on IB compared to 4.5ghz being the norm on SB.[/citation]

Due to the exponential nature of power usage and heat generation increases as clock frequencies increase linearly, we may see higher overclock differences ( by %) at lower cooling systems than in higher ones. For example, instead of a 15% improvement like on liquid nitrogen, it might be more like a 20% improvement on air.
 
[citation][nom]scook9[/nom]You demonstrate a noted lack of knowledge on how overclocking these chips works.....you barely touch the base clock, the majority of the work is done my manipulating the turbo multis well beyond what is stock, far nullifying any impact on stock settings.[/citation]
... *facepalm*
 
Status
Not open for further replies.