Sandy Bridge Reviews Released!

Ruiner said:
I have one question for anyone who has pored over all these reviews:
How much faster is a stock 2500k than an OCed 9550 (basically the best 775 has to offer) in gaming at mainstream resolutions (single gpu, ~16x12)?

Anand had a stock clock q6600 in their charts. [H] had no 775 at all and TR used a stock q9400.
Click to expand...

I noticed alot of review sites seem to include the q6600, but they never want to include the q9550\q9650 for some reason. Hopefully more reviews will come out with the q9550 in it.
 
leh18621 said:
I noticed alot of review sites seem to include the q6600, but they never want to include the q9550\q9650 for some reason. Hopefully more reviews will come out with the q9550 in it.
Click to expand...

Well clock for clock we should be about 10% faster than a Q6600, the problem is Q6600's generally don't clock anywhere near as high as a Q9550.

Still, the SB chips have such a commanding lead over the C2Q's it may be a bit of a moot point.
 
pxc said:
Comparing your specific config vs a stock 2500k is easy if the reviewer tested game(s) you play. Just benchmark your own system and compare it to those numbers in the same game(s).

Unfortunately, no one can hand you the specific benchmark numbers for your system on a platter, given the wide array of speed people overclock to and various video cards used.

Not trying to sound rude, but 1/2 the data you want is available and only you have the meaningful other half to answer your question. There are several utilities that make benchmarking easier. guru3d has several.
Click to expand...

I'll buy that, however there are several enthusiast staples that are common enough to be mentioned in these circles....just look at user sigs. Overclocked e8400, q6600 and q9550s are the staple installed Intel enthusiast base for a potential SB upgrade. Filling charts with competing i3/i5/i7 parts is nice, but I'll go out on a limb and state that most of those users won't upgrade relatively new rigs for these incremental gains.

This thread alone has over 13k views at this point, and at least a hundred readers have probably picked those charts apart enough to reply:
'Around 25%'.
 
Porter_ said:
the french review in the OP (Hardware.fr) pits SB against a lot of different CPU's - including the Q9550, Q9650, and QX9770. i assume these chips are not overclocked though.
Click to expand...

Yeah after reading the French review more in-depth, I'm not quite happy with it. Especially since I suggested adding it to the first post haha. The GTAIV and Anno 1404 were both benched at 800x600. Not the most real-world resolutions to test with, and kind of makes those charts useless imo. Ditto with the Crysis/Arma numbers. And yeah I know they said, "Le test est effectué avec les détails très élevés, mais en 800*600 afin de limiter la dépendance à la carte graphique."

For those who can't read French (and I think I'm translating correctly) - all high details, but at 800x600 to limit GPU dependence. I don't quite agree with that decision b/c nobody plays at those resolutions anymore. Sigh. They are one of the few to bench a whole slew of CPUs, I'll credit them with that effort. But using an unrealistic resolution does us no good.
 
Hell, I've been holding off upgrading my Core2 Duo at 3.2, and I have to say I'm impressed! The IPC is not much higher, but this enhanced architecture clocks incredibly high without suckuing down excessive amounts of juice. The combination of higher clock and IPC makes for one brutal 4-core chip.

For those of us who didn't jump on the Core i5/i7 train yet, now is the time! My only problem is deciding between the 2500K and the 2600K :D
 
leh18621 said:
I noticed alot of review sites seem to include the q6600, but they never want to include the q9550\q9650 for some reason. Hopefully more reviews will come out with the q9550 in it.
Click to expand...

Makes me feel like I should start a review site that actually does what the vast majority of people want. I don't see how it wouldn't succeed since nobody else is willing to do this.
 
butterfliesrpretty said:
Makes me feel like I should start a review site that actually does what the vast majority of people want. I don't see how it wouldn't succeed since nobody else is willing to do this.
Click to expand...

Go for it! I've started to get into a little IT journalism on the side covering Windows Tablet PCs, a lot of need for IT journalists that actually focus on niche stuff deep rather than being broad and generic.
 
butterfliesrpretty said:
Makes me feel like I should start a review site that actually does what the vast majority of people want. I don't see how it wouldn't succeed since nobody else is willing to do this.
Click to expand...

i imagine compiling reviews is a very time consuming endeavor so you have to be selective with which hardware you test. that or release your review a month after the fact. seems like the best way is to get your initial review out with limited hardware and then perform a followup review that's more in-depth.
 
butterfliesrpretty said:
Makes me feel like I should start a review site that actually does what the vast majority of people want. I don't see how it wouldn't succeed since nobody else is willing to do this.
Click to expand...

Good luck with that. Review sites have to use what they've been supplied or what's been supplied to them in the past. If they do not get something supplied to them in a press kit then they've got to buy it themselves. For sites which have a lot of money it's not quite as big a chore but how much money they have depends on advertising dollars. That market is fluid to the say the least. When it's up, it's really up. When it's down, it goes down like a $5 whore. Then there is space to consider. Unless they have massive offices their stock of hardware reaches a point where they have to get rid of things and they've got to pick and choose what has to go. Yeah processors don't take up too much space but boards and CPU coolers do. Additionally from what I've seen companies like Intel, AMD, and the like usually send out high end hardware to hardware review sites. They don't often send mid-range processors or low end processors even if they are enthusiast favorites. When a company needs to make an impression they send their best hardware. Not their low end or mid-range stuff.

Then there is the amount of time it takes to include one more processor in an article when it uses a different motherboard, CPU and sometimes RAM. Hell there can be quite a bit of time involved just adding an additional CPU when it uses the same testing platform. It's harder to do this stuff then you might think. Far more labor intensive than most people can imagine. The fact is you can't please everyone and you do what you can with what you've got in the time allotted.
 
Dan_D said:
Good luck with that. Review sites have to use what they've been supplied or what's been supplied to them in the past. If they do not get something supplied to them in a press kit then they've got to buy it themselves. For sites which have a lot of money it's not quite as big a chore but how much money they have depends on advertising dollars. That market is fluid to the say the least. When it's up, it's really up. When it's down, it goes down like a $5 whore. Then there is space to consider. Unless they have massive offices their stock of hardware reaches a point where they have to get rid of things and they've got to pick and choose what has to go. Yeah processors don't take up too much space but boards and CPU coolers do. Additionally from what I've seen companies like Intel, AMD, and the like usually send out high end hardware to hardware review sites. They don't often send mid-range processors or low end processors even if they are enthusiast favorites. When a company needs to make an impression they send their best hardware. Not their low end or mid-range stuff.

Then there is the amount of time it takes to include one more processor in an article when it uses a different motherboard, CPU and sometimes RAM. Hell there can be quite a bit of time involved just adding an additional CPU when it uses the same testing platform. It's harder to do this stuff then you might think. Far more labor intensive than most people can imagine. The fact is you can't please everyone and you do what you can with what you've got in the time allotted.
Click to expand...


I agree it would be and I'm not downplaying the difficulty or time, but it is getting old seeing people asking for the same few things and no review site is willing to do it.

My personal feeling on what I want to see; numbers. Generally speaking I want to see how the cpu that's coming out is compared to the popular cpus of the time at resolutions I play at.

That is by far the most important thing I want to see, and nobody is doing it.

Second to that I want to see if there's a difference in sli performance like we saw between the core 2 quads and i7's.

When all these review sites rehash the same stuff, I'm left wondering why so many of them exist. Honestly I wouldn't even mind no words at all in a review, just benchmarks.
 
butterfliesrpretty said:
I agree it would be and I'm not downplaying the difficulty or time, but it is getting old seeing people asking for the same few things and no review site is willing to do it.
Click to expand...

I just explained the reasons for this.

butterfliesrpretty said:
My personal feeling on what I want to see; numbers. Generally speaking I want to see how the cpu that's coming out is compared to the popular cpus of the time at resolutions I play at.

That is by far the most important thing I want to see, and nobody is doing it.

Second to that I want to see if there's a difference in sli performance like we saw between the core 2 quads and i7's.
Click to expand...

I don't think you understand what's going on when you look at these performance numbers. I can take a Core i7 and duplicate the results of the entire Core i7 9xx series excluding the six core 980X. I can also take the 980X and I can disable two cores and run it like any of the rest of the family. I can do the same thing with the 800 series Core i7's or any other CPUs for that matter. As long as the feature set is the same the numbers are comparable. As for older CPUs, it's up to you to do some math and do a little thinking of your own. You can go back to the articles where the Core i7's were initially compared to the Q6600 and figure out how much faster the modern processors are compared to the older stuff. You don't really need us to regurgitate information that's already out there about processors that haven't been new in almost three years. It's easy to say that a Core i7 is x% faster than the Core 2 Quad's are. Then you can look at the Sandy Bridge data and see how it compares to Nehalemm Gulftown, Westmere and more. If Sandy Bridge is x% faster than you can add the two numbers up and figure where your at since direct comparisons aren't necessarily available comparing the Q6600 to Core i7 920's and Core i7 2600K's at 4.0GHz in Metro 2033. The reason being that Metro 2033 wasn't available when the Core 2 Quad Q6600 was first released or even when it became popular.

As far as gaming performance goes the CPU isn't the most important factor anyway. This gets proven time and time again when CPU scaling articles get done. Sure there is some quantitative information showing improvements but most of those at certain points are so small that you wouldn't notice anyway. Games are most often GPU bound rather than CPU bound. You brought up SLI on Core i7 vs. Core 2. That data shows you need extremely high resolutions and multiple graphics cards to realize the performance potential of Core i7's which need to be clocked at 3.6GHz+. If you aren't running large displays or an array of displays, then I'm not sure how such information would be helpful to you. The question about Sandy Bridge vs. Nehalem / Gulftown with regard to SLI and high resolution is a valid one and I'd like to see that too. Given that many of the reviews target 2560x1600 and 1920x1200 already, I'm going to guess that your not running such displays. Furthermore when it comes to graphics cards and the resolutions "you play at." It's easy to conclude that if a card powers through games at 2560x1600 that it will perform well at 1680x1050 or 1920x1080. All it takes is a little reasoning but all the data you need is out there. Some sites do include 1680x1050 and 1920x1080 data. If it's the effect gaming on Sandy Bridge will have at "normal resolutions" with mid-range graphics cards I'm going to go out on a limb and say the effect will be limited. When hardware sites update their platforms and start doing video card articles on them you'll get a more definitive answer but we already know that CPU performance in regard to clock speed and even IPC doesn't generally translate into gaming performance. Hence the comments about games being more GPU bound.

We try to provide the most information possible but at some point reviewers have limited time and limited resources in most cases. And again the information you need is out there. At some point you have to pull the data together yourself and make your own judgement call.

butterfliesrpretty said:
When all these review sites rehash the same stuff, I'm left wondering why so many of them exist. Honestly I wouldn't even mind no words at all in a review, just benchmarks.
Click to expand...

And what do the benchmarks tell you? At 640x480 all processors haul ass? What difference does it make if a processor can give you 300FPS or 500FPS at low resolutions? At some point it will make no difference to your eyes or your 60Hz or even 120Hz LCDs. You have to pair a graphics card with your CPU and system in order to get anything accomplished anyway. And at that point there is plenty of data about systems running at 3.6GHz or whatever at 1680x1050 or whatever resolution you use. And as the [H] has proven gaming benchmarks and actual game play experiences do not always match up. I'd argue that the words, impressions, and experiences conveyed by the author of the articles is worth far more than the canned benchmarks are. (Benchmarks which are by nature designed to isolate specific components in unrealistic scenarios.)

The bottom line is that if you want to know how your old hardware stacks up, your going to have to fire up some tests of your own or use a search engine to help you figure it out. I think what you are looking for goes beyond getting an answer to question. I think what you are looking for is someone to do your thinking for you.
 
Last edited:
Dan_D said:
Good luck with that. Review sites have to use what they've been supplied or what's been supplied to them in the past. If they do not get something supplied to them in a press kit then they've got to buy it themselves. For sites which have a lot of money it's not quite as big a chore but how much money they have depends on advertising dollars. That market is fluid to the say the least. When it's up, it's really up. When it's down, it goes down like a $5 whore. Then there is space to consider. Unless they have massive offices their stock of hardware reaches a point where they have to get rid of things and they've got to pick and choose what has to go. Yeah processors don't take up too much space but boards and CPU coolers do. Additionally from what I've seen companies like Intel, AMD, and the like usually send out high end hardware to hardware review sites. They don't often send mid-range processors or low end processors even if they are enthusiast favorites. When a company needs to make an impression they send their best hardware. Not their low end or mid-range stuff.

Then there is the amount of time it takes to include one more processor in an article when it uses a different motherboard, CPU and sometimes RAM. Hell there can be quite a bit of time involved just adding an additional CPU when it uses the same testing platform. It's harder to do this stuff then you might think. Far more labor intensive than most people can imagine. The fact is you can't please everyone and you do what you can with what you've got in the time allotted.
Click to expand...

Let's not exaggerate Q6600+Q9550+E8400+mobo+ram is <1000 euro of equipment.

Cost of buying i3 530 and then reselling it would be like 20 euro or even less hardly important compared to price of manpower to do those tests.
 
I've never had trouble figuring out what's what with hardware, after reading a review here plus one or two other review sites. They all can't meet everyone's needs, but it's never difficult to figure out what the given results mean. And as it's been two years since upgrading for me - I too come to the conclusion mentioned above that EVERYTHING mid-level and up hauls ass out there now.
 
Michaelius said:
Let's not exaggerate Q6600+Q9550+E8400+mobo+ram is <1000 euro of equipment.

Cost of buying i3 530 and then reselling it would be like 20 euro or even less hardly important compared to price of manpower to do those tests.
Click to expand...

You assume that the hardware will survive the testing and will be in good enough condition to resell. Also you tend to lose more than just a few bucks reselling hardware. Some people have good luck sticking it to people for nearly full price but I wouldn't bet on that. Not when you can go to the store or online and get the same product new with the ability to exchange it for nearly the same price. Don't forget if a site doesn't have a Q6600 or whatever on hand, they've got to procure one and at the start of a new generation like we see now with Sandy Bridge's release pricing plummets on some older stuff. So after the completion of the article there is a good chance that the hardware you bought will be worth substantially less once the article is released. That's not always the cause but it certainly can factor into the situation. Also shipping and man hours factor into it. Labor isn't free. Adding just one more processor or one more platform into the mix can add several hours to the process. Especially if you run into issues for any reason.

Anyway, we are starting to stray off topic. This is about Sandy Bridge and much less about aging hardware and why it wasn't included in an article.
 
Dan_D said:
I just explained the reasons for this.



I don't think you understand what's going on when you look at these performance numbers. I can take a Core i7 and duplicate the results of the entire Core i7 9xx series excluding the six core 980X. I can also take the 980X and I can disable two cores and run it like any of the rest of the family. I can do the same thing with the 800 series Core i7's or any other CPUs for that matter .
Click to expand...

I've seen alot of sites do this and I don't believe the numbers will be accurate sure you can disable cores on the 980x but what about the cache? that is still a 12mb cache chip vs 8mb L3 cache.

It like reviews i've seen when they disable 2 cores on a i7 and then say this is what you will see with their E6400 or whatever they have which is wrong because a dual core i7 should be faster than a dual core Core2 or athlon or whatever the comparison is.
 
Makaveli@BETA said:
I've seen alot of sites do this and I don't believe the numbers will be accurate sure you can disable cores on the 980x but what about the cache? that is still a 12mb cache chip vs 8mb L3 cache.

It like reviews i've seen when they disable 2 cores on a i7 and then say this is what you will see with their E6400 or whatever they have which is wrong because a dual core i7 should be faster than a dual core Core2 or athlon or whatever the comparison is.
Click to expand...

I meant within a given family. I'm not talking about using a Core i7 to simulate LGA1156 chips or Core 2 Quad's. It's true that the cache is larger and I hadn't thought about that. I was making a point more than anything. With one higher end processor you can simulate the rest. Typically just by reducing the multiplier. If I took a Core 2 Extreme 965 for example, I could simulate the 920, 930, 945, 950, etc.
 
Sweet. I wonder if they'll have decent mobos to go with them. ~20min drive for me.

edit: unless newegg can beat that when tax is rolled in
 
Last edited:
Will the 6 vs. 8 meg L3 cache really show any real world improvements? What I'm gettin at is; would you even notice buying the 2500k and saving $100?
 
Damn, thats nice pricing, what is the typical per customer limit? If I had access to MC I would probably buy a few, keep the best ocing chip and sell the rest on FS/FT (disclosing what I got for OC) or ebay. I think you could probably break even on the extra chips and ensure you get a good one, but thats just me.
 
kooda said:
Will the 6 vs. 8 meg L3 cache really show any real world improvements? What I'm gettin at is; would you even notice buying the 2500k and saving $100?
Click to expand...

from a gaming standpoint it it doesn't have much effect at all from the reviews i've read. you might want to go through some of the reviews in the OP to better draw your own conclusions though.
 
From what i've seen the extra cache really seems to help the chip with HT. If its worth $100 is up to you.
 
kooda said:
Will the 6 vs. 8 meg L3 cache really show any real world improvements? What I'm gettin at is; would you even notice buying the 2500k and saving $100?
Click to expand...

It seems like the 2600s get a little better overclocking results, but only a little, like 1 or 2 multipliers. I was set on a 2600, but now I'm thinking 2500 instead.
 
I have been running a 2600k for a few days now and i am very impressed with the performance. I have had I5 750, i7 860 and 1055T all at 4.1-4.3 ghz and this 2600k just smashes all of them especially since mine will do 4.7 easily. Especially with video encoding. Run x264 bench..... just wow.....
 
kirbster said:
I have been running a 2600k for a few days now and i am very impressed with the performance. I have had I5 750, i7 860 and 1055T all at 4.1-4.3 ghz and this 2600k just smashes all of them especially since mine will do 4.7 easily. Especially with video encoding. Run x264 bench..... just wow.....
Click to expand...

nice to see 4.7 GHz while using a moderate cpu cooler. how much voltage did you need and what kind of temps are you getting under load?
 
I have a high output fan on the Zalman, but i dont think it is necessary. 4.7ghz @1.38v i never see temps even hit 60C under full load.

I think this is a really underrated cooler as it took my 1055T up to 4.5ghz as well.
 
Dan_D said:
snip.
Click to expand...

I agree with some of what you're saying, but you have to really think about why people are reading these reviews. I think most of the time it's because they are wanting to see if this new cpu is a worthwhile purchase for their needs.

One point I would like to make is that it's very hard to find certain specific q9550 benchmarks at all. Strangely hard, especially ones that compare current games with current video cards against an i7, let alone the new sandy bridge cpus @ 1080p. A lot of us have this chip and want to know if it's a worthwhile upgrade..... that's it, but all these reviews leave us scratching our heads.

Another thing to throw into the mix is that different cpus will perform differently depending on what in game settings you use.
 
Last edited:
butterfliesrpretty said:
I agree with some of what you're saying, but you have to really think about why people are reading these reviews. I think most of the time it's because they are wanting to see if this new cpu is a worthwhile purchase for their needs.

One point I would like to make is that it's very hard to find certain specific q9550 benchmarks at all. Strangely hard, especially ones that compare current games with current video cards against an i7, let alone the new sandy bridge cpus @ 1080p. A lot of us have this chip and want to know if it's a worthwhile upgrade..... that's it, but all these reviews leave us scratching our heads.

Another thing to throw into the mix is that different cpus will perform differently depending on what in game settings you use.
Click to expand...
i hear what you're saying. benchmarking suites change all the time, so comparison data is hard to find. This is often an issue with gpu reviews. I couldn't find many, if any, that used consistent test methods when I wanted to compare the generation-old 4870 (which I had) to the gtx 460. Anyway, I see Dan D's point. If you want to compare your stuff to their stuff, just run your own benchmarks and see how they compare. It won't be perfect since you'll probably have a number of hardware differences but it will be a ballpark estimate. expecting reviewers to include a full spectrum of stuff more than a generation old in reviews is a little unreasonable.

either way, a 2600k might be my next dedicated folding rig.. one of those babies at 4.8ghz doing bigadv sounds like a budget way to jump in and get respectably close to what the big boys are pumping out with 980x's
 
You must log in or register to reply here.
Back
Top