You are currently viewing our boards as a guest, which gives you limited access to view most discussions and access our other features. By joining our free community, you will have access to post topics, communicate privately with other members (PM), respond to polls, upload content, and access many other special features. Registration is fast, simple, and absolutely free, so please, <a href="/profile.php?mode=register">join our community today</a>!
You should have no problems providing some forum names of people who did this then if it is as common as you allege.
Of course if it wasn't so common, you won't be able to provide any names.
Forum names?Are you kidding?I was referring to general sentiments that were echoed on various "big" forums,some of them being XS,AT forum,THG forum,OC3D etc. Ace's is(no offense) the "niche" one,having rather small member base.And even with such a small member base,if i had time to waste on search,I bet i could find some posts that fit right into my description.
Well, I do not visit these other forums. Thus, for people like me, it would be nice to either back-up what you say based on this forum, or - much better - just refrain from stating it in the first place.
Sorry if I hurt your feelings man :(.I won't refrain from stating my opinion.I based it on my personal experience so it may be different for you. The fact that you didn't "see it" does not mean it didn't happen(maybe not on this small(ish) forum).
I agree. Besides physics calculation, as good as all FP calculations have been moved to the graphics cards or am I forgetting something? A decent game spends about 20 to 30% in AI routines (purely integer AFAIK), world calculations and other position calculations are also purely integer.
Typically games are very sensitive to the average cache/memory latency (so the L2-cache latency and size has a rather important role here as the L1 of the CPUs are pretty close if we exclude the P4)
I based it on my personal experience so it may be different for you. The fact that you didn't "see it" does not mean it didn't happen(maybe not on this small(ish) forum).
You based it on your irrational sense of AMD victimhood and you tried to present some rarely seen statement as being widespread.
When called upon to show that such statements were widespread, you came up with nothing, as to be expected.
This review has only handful of tests and gaming tests are limited by a slow laptop GPU. There are many tests where Core2 is 30-50% faster than K8, it is safe to assume that K8 is about at least as fast (if not faster) than Yonah in these cases.
Now we aren't agreeing and there are no sources, I know there is a few (not so many) tests where Core 2 is 30-50% faster than K-8, but it doesn't mean Core 2 is 50% faster than Yonah, many of the tests wich Core 2 is very faster (like Photoshop or SuperPi) Yonah was fast too.
The relative speed between Conroe and Yonah doesn't vary as much as it varies between Conroe and K-8.
You based it on your irrational sense of AMD victimhood and you tried to present some rarely seen statement as being widespread.
When called upon to show that such statements were widespread, you came up with nothing, as to be expected.
Seems your feelings were hurt too :).Ah well.
EduardoS wrote:
Now we aren't agreeing and there are no sources, I know there is a few (not so many) tests where Core 2 is 30-50% faster than K-8, but it doesn't mean Core 2 is 50% faster than Yonah, many of the tests wich Core 2 is very faster (like Photoshop or SuperPi) Yonah was fast too. The relative speed between Conroe and Yonah doesn't vary as much as it varies between Conroe and K-8.
Found some oldish THG Yonah Vs Merom review here:
http://www.tomshardware.com/picturestor ... 3-ghz.html Like AT numbers,here we can see the same trend,a 0-20%.Merom part was running at higher clock in the review so that has to be accounted for .
This review has only handful of tests and gaming tests are limited by a slow laptop GPU. There are many tests where Core2 is 30-50% faster than K8, it is safe to assume that K8 is about at least as fast (if not faster) than Yonah in these cases.
K8 is faster(or on par) with Yonah on average.Yonah had a weak FPU and it was felt primarily in games. But the fact is that Merom is still on average 0-25% faster than Yonah and you must look at the applications and not just games.In AT review there are typical desktop workloads that show gains in the range EduardoS stated. So to recap,on client type workloads K8 is on par or slightly faster than Yonah and Merom is around 20-25% faster than K8(there are cases where K8 is faster or not slower than Merom,but these are rare of course).
Honestly, you're much better off using benchmarks like SPEC as a metric for performance rather than some nebulous 'desktop' games.
Obviously desktop games are important and real applications matter, but SPEC has the advantage of being a widely used and standardized benchmark.
This review has only handful of tests and gaming tests are limited by a slow laptop GPU. There are many tests where Core2 is 30-50% faster than K8, it is safe to assume that K8 is about at least as fast (if not faster) than Yonah in these cases.
K8 is faster(or on par) with Yonah on average.Yonah had a weak FPU and it was felt primarily in games. But the fact is that Merom is still on average 0-25% faster than Yonah and you must look at the applications and not just games.In AT review there are typical desktop workloads that show gains in the range EduardoS stated. So to recap,on client type workloads K8 is on par or slightly faster than Yonah and Merom is around 20-25% faster than K8(there are cases where K8 is faster or not slower than Merom,but these are rare of course).
Honestly, you're much better off using benchmarks like SPEC as a metric for performance rather than some nebulous 'desktop' games.
Obviously desktop games are important and real applications matter, but SPEC has the advantage of being a widely used and standardized benchmark. DK
If you pay closer attention,games are just one minor part of the reviews we have links for.In typical client workloads, for which these cores-Yonah and Merom- were primarily designed and targeted ,we can clearly see the 0-20% figures.Games are showing least of the gain since they are mostly GPU limited.
This review has only handful of tests and gaming tests are limited by a slow laptop GPU. There are many tests where Core2 is 30-50% faster than K8, it is safe to assume that K8 is about at least as fast (if not faster) than Yonah in these cases.
K8 is faster(or on par) with Yonah on average.Yonah had a weak FPU and it was felt primarily in games. But the fact is that Merom is still on average 0-25% faster than Yonah and you must look at the applications and not just games.In AT review there are typical desktop workloads that show gains in the range EduardoS stated. So to recap,on client type workloads K8 is on par or slightly faster than Yonah and Merom is around 20-25% faster than K8(there are cases where K8 is faster or not slower than Merom,but these are rare of course).
Honestly, you're much better off using benchmarks like SPEC as a metric for performance rather than some nebulous 'desktop' games.
Obviously desktop games are important and real applications matter, but SPEC has the advantage of being a widely used and standardized benchmark. DK
If you pay closer attention,games are just one minor part of the reviews we have links for.In typical client workloads, for which these cores-Yonah and Merom- were primarily designed and targeted ,we can clearly see the 0-20% figures.Games are showing least of the gain since they are mostly GPU limited.
You're totally missing my point. By and large I don't trust reviewers to do a good job.
I'd like to think that I do a good job, but I know that there are things I don't document, which I probably should, or settings that I just leave at default, that maybe I should play with.
Especially when you measure power, things get very complicated, very quickly and there's a lot of different things you need to pay attention to that most reviewers don't.
OTOH, I trust vendors to do an excellent job of optimizing their systems. Obviously, it may not reflect what happens to a normal system from HP bogged down with bloatware...but I expect that all the vendors are 'fudging' equally well.
Also, comparing the K8 and Yonah isn't quite sensible...as they target rather different markets. If you want to compare a notebook SKU of K8 to a Yonah, that makes sense.
But Yonah was frequency capped because it was designed for notebooks. K8 was designed for higher frequencies since it went into servers, workstations and desktops.
We used Yonah for comparison since it was a basis for Merom design.We wanted to see how much IPC did come from all the stuff intel implemented and compare that to the K8->10h transition.As can be seen from client type workloads,both netted about the same 10-20% in int and massive improvement in SSE(which is the one of the biggest changes in both).
We used Yonah for comparison since it was a basis for Merom design.We wanted to see how much IPC did come from all the stuff intel implemented and compare that to the K8->10h transition.As can be seen from client type workloads,both netted about the same 10-20% in int and massive improvement in SSE(which is the one of the biggest changes in both).
I haven't seen any analysis that is even relevant.
Also, there's a big difference between Yonah-->Merom and K8-->Barcelona.
Yonah topped out at 2.2GHz, merom hit 3GHz (in fact, you could get a Merom MCP at 3GHz IIRC). K8 topped out at 3.2GHz and Barcelona hit no more than 2.9GHz, although initially I don't seem to recall them getting much above 2.4GHz.
In other words, you don't want to focus too much on IPC because what really matters is performance and if you gain IPC by losing frequency that's bad.
I realize you are trying to somehow claim that K8-->Barcelona was as big a step as Yonah-->Merom, but frankly you're out to lunch. Besides Barcelona's contemporary was Penryn anyway...
Just compare phenom II and last core2 and compare с2d vs that time amd chips. You see diffference is much smaller. But nehalem "saves the day".
SSE was greatly improved in phenom, it used by a lot multimedia programs, codecs etc.
So that's the second die photo of Sandy Bridge, a different photo of the same die. Probably the correct one then :^)
Conclusion:
Seems highly likely now that Sandy Bridge's 256 bit AVX is implemented with 128 bit hardware running at twice the frequency, and becoming likely now: That Larrabee's 512 bit LNI uses the "same" 128 bit hardware running at 4 times the clock speed as the rest of the core....
in the 2nd bullet he says : "[...] measures 1.42X speedup (would have predicted 1.5X with the current architecture in the limit; vs. 1.0X if we had double pumped). [...]"
so he is explicitely stating that it isn't double pumped, isn't it ?
Users browsing this forum: No registered users and 2 guests
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot post attachments in this forum
