ATI HD 5870 Scaling Performance: X58 vs P55 Showdown
by Gary Key on September 29, 2009 12:00 PM EST- Posted in
- Motherboards
We will start off our P55 coverage this week by answering a question that has been raised numerous times as of late, “Exactly how will the new DX11 cards perform on the Lynnfield/P55 platform compared to the Bloomfield/X58?”. The answer to that question depends on the game engine, settings, and processor choice for the most part. There has been much speculation that the Lynnfield/P55 platform would fail miserably with the next generation cards. That said, the difference in platform performance between the first DX11 capable cards available (ATI HD 5870/5850) is about the same as previous generation cards we tested in the Lynnfield launch article.
This means for single card performance both platforms trade blows for supremacy. However, for those running CF/SLI setups, the X58 continues to be the platform of choice for users wanting the best possible benchmark results. Does that mean the integrated dual x8 PCIe 2.0 logic on Lynnfield is a poor choice compared to the dual x16 PCIe 2.0 sporting X58? Absolutely not based on our initial tests. In fact, it should satisfy most users.
Now for those making an investment into an ultra high-end HD 5870 CrossFireX setup, the Core i7/X58 configuration will simply offer the best possible performance. Of course that performance comes at a cost, particularly power consumption. However, we have a feeling most owners sporting two HD 5870s are not that concerned about Al Gore knocking on their door in the middle of the night.
In the meantime, we have our first performance results comparing the Core i7/860 and P55 against the like priced Core i7/920 and X58 in a variety of games. We are going to state this upfront, this is not a GPU review of the HD 5870. Instead, we decided to pull this information out of the upcoming high-end P55 roundup so it did not get lost in the mix. Our resolution is limited to 1920x1080 that we use in the motherboard test suite. As such, the numbers speak for themselves. So let’s get right to the results today, but first, the test setup.
Test Setup-
For our test results we setup each board as closely as possible in regards to memory timings and sub-timings. The P55 motherboard utilized 8GB of DDR3, while the X58 platform contained 6GB. The P55 and X58 DDR3 timings were set to 7-7-7-20 1T at DDR3-1600 for the i7/920 and i7/860 processors at both stock and overclocked CPU settings. All power management features were enabled on each board and voltages were set at the lowest possible values when overclocking while still retaining 24/7 stability.
The image gallery below contains our Everest memory results with each processor overclocked at similar memory settings along with voltage/uncore/subtiming options. The 860/P55 offers slightly better throughput and overall latency numbers than the 920/X58 when overclocked. At stock, the latency numbers favor Lynnfield with assistance from the turbo mode.
Our game selection today is varied but is missing a couple of titles we wanted to show. We pulled Crysis Warhead as the CrossFireX scaling numbers were under 8% and Need For Speed: Shift does not have a CrossFire profile yet. AMD is currently working on updated CF profiles for the latest games. We also had some corruption problems when running Empire:Total War and Anno 1404 that is under investigation.
All of our gaming performance results are reported in average frame rates per second in the main bar graph with minimum frame rates reported in the text section. The results are sorted by the average frame rates. We are also reporting single card results with the HD 5870 running at PCIe 2.0 x8 speeds on the P55 platform to compare performance to the x16 single card setup. We installed an Intel CT Gigabit network card in the second physical x16 slot in order to force x8 operation.
Facebook
Twitter
RSS
85 Comments
View All Comments
TimboG - Wednesday, September 30, 2009 - link
"If you actually owned either system you would know a 200MHz uncore clock improvement does nothing for the benchmarks."That's where I have a problem.
We read these articles BEFORE making a purchase to aid us in making the correct choice for each of our needs. I felt that if these benchmarks were to be done as a comparison of platforms then there should be much fewer results to be published.
Consider a comparison of two platforms.
1. Platform A with single GPU
2. Platform B with single GPU
3. Platform A in CF/SLI
4. Platform B in CF/SLI
No CPU, Mainboard, or GPU settings changes.
K.I.S.S.
Yes, He has those results, maybe, but how do I know. Maybe he overlooked something when making changes back and forth between settings.
Not to say he did, but we are all human. Testing these platforms at their default settings is a TRUE comparison.
Who cares if there is a difference in CPU speeds at stock settings? Is this not what we will get out of the box? That is the difference in platforms. If you want to compare CPU differences then add more results with more STOCK CPUs added to the test with each platform.
ggathagan - Wednesday, September 30, 2009 - link
What exactly is your beef?He DID exactly that:
1. Platform A with single GPU
i920/HD 5870
2. Platform B with single GPU
i860/HD 5870
3. Platform A in CF/SLI
i920/HD 5870 CF
4. Platform B in CF/SLI
i860/HD 5870 CF
As a bonus, along with the stock CPU, Mainboard, and GPU settings that you crave, he also made an effort to minimize the differences between the two CPU's by matching their clock speeds at 4.2 Ghz.
After all, the question he was trying to answer is: Does it make a difference whether you use the X58 or the P55?
He even went one better: If you force the P55 into x8 PCIe mode with a single GPU, does it make a significant impact on performance?
Perhaps reading more closely would be helpful to you:
The P55 and X58 DDR3 timings were set to 7-7-7-20 1T at DDR3-1600 for the i7/920 and i7/860 processors at both stock and overclocked CPU settings.
That's not changing settings; both are using the same memory and same memory timings.
All power management features were enabled on each board and voltages were set at the lowest possible values when overclocking.
The only time settings were altered from their default was when overclocking.
Nowhere in the article is GPU overclocking mentioned.
TA152H - Tuesday, September 29, 2009 - link
Are you really that stupid?Nevermind, I already know the answer.
The uncore on the Lynnfield is 3602. The Bloomfield 3406. Are you seriously that stupid that you think that doesn't have any performance impact? It has an effect on L3 cache speed as well as memory. Are you implying he can't get closer to 3602 than 3406? It's just not apples to apples. Compare apples to apples, and you get a bigger difference between the two. Not that they'd want to show that.
Overall, I agree with him the Bloomfield offers better performance, and that the brain-damaged Lynnfield is attractive in areas where you don't need absolute performance and the power characteristics are attractive. But, I hate the testing methodology. Originally, they were trying to convince us the Lynnfield wasn't any slower. Thank goodness they gave up that nonsense.
mesiah - Wednesday, September 30, 2009 - link
TA, for the longest time I couldn't figure out why you weren't banned from this site. Then I realized that they keep you around for pure entertainment value. But it really is getting old. I can pretty much sum up every one of your posts in a few short sentences:@the person above me, You are a moron, you aren't half as smart as I am so don't even bother giving on opinion. Here is a bunch of old random facts that I will puke out to make myself feel superior. Also, lynnfield is brain dead and intel does not know how to market processors. Gary, you suck, my dog writes better than you, and does a more thorough job of research. Anyone who disagrees with me is a complete idiot, and so on and so forth.
TA, please feel free to cut and paste my above paragraph into any of your future responses. It will save you some time.
goinginstyle - Wednesday, September 30, 2009 - link
"The uncore on the Lynnfield is 3602. The Bloomfield 3406. Are you seriously that stupid that you think that doesn't have any performance impact? It has an effect on L3 cache speed as well as memory. Are you implying he can't get closer to 3602 than 3406? It's just not apples to apples. Compare apples to apples, and you get a bigger difference between the two. Not that they'd want to show that. "I own both platforms and just ran the FarCry 2 bench they use but with a 4870x2. It makes no difference dude. If you think it does, then let's see your proof. The fact they left the uncore ratios at stock settings is apples to apples. Why do you want them changing settings so bad. So you can bitch about them changing settings. It is so obvious what you are trying to do here.
Gary stated in the last article that the uncore makes no difference at these speeds. It does not and looking around the web this evening it appears only uncore speeds above 4GHz make a difference and that is with the extreme guys on LN2 running things like superpi and 3dmarks.
Where is your article showing otherwise? When will it be posted at your site? Why do you not comment at Toms in the same way you do here. They have not provided this information and if they do I bet you do not comment on it. Why is that?
brybir - Tuesday, September 29, 2009 - link
Why do all of your posts imply that Intel has absolutely no idea how to market segment its products or properly utilize its production, research and fabrication facilities to their maximum potential?Most of your post is just speculation on your part and based on no actual evidence of anything. You do not know what Intel's 32nm roadmap is, what products will definitely be produced etc, and you certainly do not know how they will sell and how they will be priced.
And your comments about lobotomized processors is just nonsense. You seem to believe, based on technical shortcomings that you seem to think are important but do not translate into real world difference, that the i7/i5 generation is somehow crippled. Yet, benchmark after benchmark show you to be wrong, with the i7 platform on par with the i9 platform except in the areas of extreme high end systems like in this review. Additionally, your comments about market segmentation are equally nonsense. You do not have Intel's sales data. You do not work for Intel in a position of responsibility for shaping corporate decisions. Intel will segment in the way that makes them the most money. Your opinion as to what they "should be" are based on nothing more than your opinion unsupported by any actual facts.
In any event, you seem to crap on every review by Gary and the Anandtech staff based on these obscure technical issues that you think are important but have been shown not to be. Instead of coming in here and insulting the excellent work done by Gary and his staff, perhaps you could ask the question in a more constructive way and actually get one of them to respond to you as they do many others. Let me give you some examples:
1. Your Statement: "Another bad review by Gary. I was hoping you'd have learned something about testing by now, but, apparently not. For the rest of you, you'll notice Gary is up to his old tricks. He clocks..."
Better way of saying it: "Gary, thank you for taking the time to work on this review, I found the information to be interesting, but I do have some questions about some technical settings that you used specifically relating to your choice of uncore settings and the impact that would have on these benchmarks given its potential impact on memory access, latency and speed. Perhaps you could explain what your thinking was so that I better understand your decisions and why you made them"
Your way makes you sound like an arrogant jerk coming in here only to piss in the lemonade of everyone that enjoys coming to this site to talk hobby shop and discuss the article. Try some tack. Drop the attitude. Some of your arguments have some merit to them (or may have merit, I do not know), but most of us have summarily stopped reading your posts in the same way that we ignore the crazy guy on the corner preaching the end of the world every day for five years.
So, unlike the others, I want you to stay here. Make your points. Engage in discussion. Just drop the attitude, arrogance and derisions in your posts and try to be a decent member of this community. You have interesting things to say sometimes but your method of communicating makes anything useful you have to say fall on deaf ears.
TA152H - Wednesday, September 30, 2009 - link
Wow, I'm being patronized by someone with half my IQ. And not even well.That's actually the worst attitude. That patronizing attitude like you're superior, and you're so kind as to want me to post. Should I be grateful for your magnanimity? How hypocritical.
Now, more to point. Intel did market this stuff wrong. Their sales say it. I only predicted it, and bring it up to prove a point. They're not omniscient. It's a fact that Intel does not use their resources perfectly, as no company possibly can. It's only a matter of degree of imperfection, and they're seeing some pretty good level of imperfection with their marketing right now.
Your line about their 32nm roadmap proves you're a complete moron and can't read. That's exactly my point. They're showing too few products, and must have something else unless they can sell the Clarkdale really well.
Again, you're misrepresenting me when you say I claim they can't make marketing decisions at all. You obviously shouldn't comment when you can't understand any of what I'm saying. I said putting the MMU on the GPU is BAD from a technical perspective. I also said I expect it will sell very, very well a number of times. Don't simplify what I say so you can understand it.
You're also simplifying when you call it the i7. I think the i7 is good; the Bloomfield version. Benchmark after Benchmark do not show that the Lynnfield is equal to the Bloomfield. Even with Gary skewing the results, the Bloomfield wins. How can you read this and still not see that?
Before you start patronizing, actually try to have a clue what you're talking about. Having vague notions without any substantive understanding doesn't quantify as proof.
And, if you noticed, I've at least got them to back off their horsecrap about Lynnfield having equal performance, and the PCIe being a better, faster implementation. That was making me crazy, it was so false. I can't expect them to admit they're wrong, because they're not, and I'm not. It's opinion about which is better, and also situational. But, saying it's as fast, and has a better PCIe implementation is just plain false. I can accept the current opinions, especially since he's dead on with the x58 being way too power hungry. It's a problem they should have taken care of a year later.
slurmsmackenzie - Wednesday, September 30, 2009 - link
wats wrong wit sum dummerd critizing u wen hes rite. ur an ass. even a moron as you call it can see that. your big words were delivered like an ass. don't need to know the meaning of magnan..mag..whatever to know an ass said it.goinginstyle - Wednesday, September 30, 2009 - link
TA152H..."Wow, I'm being patronized by someone with half my IQ. And not even well. "
Man, do you ever have a high opinion of yourself. I guess being middle aged and still living with your mother will give you one hell of a superiority complex.
"You're also simplifying when you call it the i7. I think the i7 is good; the Bloomfield version. Benchmark after Benchmark do not show that the Lynnfield is equal to the Bloomfield. Even with Gary skewing the results, the Bloomfield wins. How can you read this and still not see that? "
Man, what benches have you been looking at or do you just imagine this stuff up. In applications it is Lynnfield ahead in just about every bench and in gaming it holds true with the nvidia cards. The ati cards produce a different result and that is about it. All of the review sites have the same results. They are not skewed here and apparently you have a personal vendetta against Gary. Once again, why do you not post at the other sites who have the same conclusions? What is your motive for spamming every article by Gary or Anand? Where is your P55 article dude?
"And, if you noticed, I've at least got them to back off their horsecrap about Lynnfield having equal performance, and the PCIe being a better, faster implementation. That was making me crazy, it was so false."
They have not backed off anything, their message from the first article to this last one has been the same. Apparently you have a hard time reading. Maybe you should get your mother to read the articles for you when she tucks you in.
Inkie - Wednesday, September 30, 2009 - link
"I said putting the MMU on the GPU is BAD from a technical perspective."Well, tell that to the Intel processor designers responsible for Clarkdale and they will laugh at you.
"You're also simplifying when you call it the i7. I think the i7 is good; the Bloomfield version. Benchmark after Benchmark do not show that the Lynnfield is equal to the Bloomfield. Even with Gary skewing the results, the Bloomfield wins. How can you read this and still not see that?"
In Anand's review he clearly said that X58 can be superior for CF/SLI.