Oh so some people at Best Buy thinks it's a real core so you assume everyone else would think that too? Like the others have said real CPU cores takes more space and more power while HT only requires a little more power and a little more space. And not to mention HT is one of the reasons why i7s are so effective against Thubans in highly threaded benches.Originally Posted by stangracin3
You would lose that bet. Along with multiple threads from my browser, there are >150 processes (some with their own multiple threads) running as I'm taking a moment to type this message. Many are idle system processes, but they could wake up and need processor time. Some are doing real work. If I had more cores I could do even more work, be it running more threads from a small number of programs or running more single-threaded programs simultaneously.I bet the browser you using to type that message is single threaded.Although I know you meant a single threaded app that maxes out the cpu
Exactly. Multitasking has been around a long time. Though it doesn't translate to benchmarks well, it's an important usage scenario for many users.Your last sentence is key there. The vast majority of apps are single threaded bound but end users will like been able to run multiple apps at once without them affecting each other, multi core is defenitly useful for that.
solus what browser? IE eg. has multiple processes but it is not multiple threaded. Firefox and opera are also both not multi threaded. Flash is not multi threaded.
Isn't google chrome multithreaded?
Each browser window, is it's own thread though... right?
Each tab (process) open in IE is 18-30 threads according to resmon
My firefox has 27 threads running. Though I realize you probably meant a thread per tab ( I don't know if it's in this nightly yet ). But even if all the tabs are a single thread, you'd still have separate threads for user interaction, networking, etc. A browser will usually be network limited before CPU limited anyway, except on low power processors. Separate threads for each tab would help on dual core atom.
So I have heard. And IE8 as well?
it runs a different process for each tab. that's to prevent one tab from crashing another which could happen with using a thread for each tab.
Welcome to intel retail edge program. A couple percent of informed users that understand the value of real cores over virtual cores doesn't add a lot to the bottom line. The people that walk into best buy aren't "some" people, rather "a lot" of people, and to them the sales people have the power over what they buy.
As for the power debate didn't RichieRich just show that 6 real cores use the same power as 4 real plus 4 imaginary cores? And then it was established that its not right to compare that way?
You're right, so why don't we compare Intel's 6 real cores plus 6 virtual cores to AMD's 6 real cores?
His argument was that because Thuban (1090T) was a new design, you can't exactly compare it to Nehalem (i7 930) which is basically roughly 10 months old. Just like how it's not right to compare 1090T to a 965BE.
Whaaat?? That doesn't make any sense at all, why would anyone say that. Websites make those comparisons daily in their reviews. Especially the intel-centric sites like anandtech, as long as it works in intels favor. some might have a disclaimer of a couple obscure words in brackets aluding to some aspect of the fact, but the graphs are shown, and the data is presented. So the comparisons are made routinely. /You're right, so why don't we compare Intel's 6 real cores plus 6 virtual cores to AMD's 6 real cores?
His argument was that because Thuban (1090T) was a new design, you can't exactly compare it to Nehalem (i7 930) which is basically roughly 10 months old. Just like how it's not right to compare 1090T to a 965BE.
there is no "virtual core". it's one core (and a fast one at that) running two threads pseudo-simultaneously. generally people who dont like HT are people who dont have it. i find this interesting and coincidentally amd's future uarch does HT* so much better. you'd better hope that's true considering nehalem has been around since 2008 and BD is 2011.
* hyperthreading is a trademark of intel corporation not to be confused with SMT
why you guys are even comparing it to the 9xx seriers... the 8xx seriers has the same performace, and consumes less (platform wise) and is even the same price range as thuban....
I myself am not a fan of the comparison, and I think it's perfectly valid to compare the power consumption of the those different processors.
But I can understand where he's coming from.
i agree.maximumpc did a quick review and the i7-800 series is very competitive vs the 1090T.
AMD's $300 6-Core CPU: Too Good to Be True?
and tech report did a article on the new 875K showing it not only keeps up but is a better performer
and even is a better value on a platform cost and overall power consumption.
The new value mix-techreport
Last edited by radaja; 06-26-2010 at 10:19 AM.
_________________________________________________
............................ImAcOmPuTeRsPoNgE............................
[SIGPIC][/SIGPIC]
MY HEATWARE 76-0-0
I know the stupidest thing I have ever done with chrome is before going on a vacation I preloaded 100 youtube videos of a tv series on my laptop so I wasn't bored during the drive. Yea not a good idea to load 100 Chrome tabs ever.
My rig the Kill-Jacker
CPU: AMD Phenom II 1055T 3.82GHz
Mobo: ASUS Crosshair IV Extreme
Game GPU: EVGA GTX580
Secondary GPU 2: EVGA GTX470
Memory: Mushkin DDR3 1600 Ridgeback 8GB
PSU: Silverstone SST-ST1000-P
HDD: WD 250GB Blue 7200RPM
HDD2: WD 1TB Blue 7200RPM
CPU Cooler: TRUE120 Rev. B Pull
Case: Antec 1200
FAH Tracker V2 Project Site
Yup, no tab overflow and the fact that Chrome uses quite a bit of memory. But that latter part is pretty much a result of each tab being its own process.
Microsoft even had a concept that went way further than what Chrome did. Microsoft's concept gave every single page element that came from a different source its own process. This would mean that each different ad banner on a page would have it's own process for example, as each ad usually comes from a different source.
"When in doubt, C-4!" -- Jamie Hyneman
Silverstone TJ-09 Case | Seasonic X-750 PSU | Intel Core i5 750 CPU | ASUS P7P55D PRO Mobo | OCZ 4GB DDR3 RAM | ATI Radeon 5850 GPU | Intel X-25M 80GB SSD | WD 2TB HDD | Windows 7 x64 | NEC EA23WMi 23" Monitor |Auzentech X-Fi Forte Soundcard | Creative T3 2.1 Speakers | AudioTechnica AD900 Headphone |
Saw that review too, but how can they say it's a better value on platform cost? Can't use DDR2 or cheaper mobos with the i7-875k. Anyways I'll stick back to the topic.
This will probably just be a tease and nothing really special of course. Just being pessimistic.
silicon available earlier then predicted???would be fun
WILL CUDDLE FOR FOOD
Except when it comes to the benchmarks there would be complaints of DDR3 vs. DDR2. Anyway you slice the pie, there would be complaints.
ok I will try to make what I am saying clearer as its causing confusion.
if I load a page thats demanding on the cpu, it will do one of 2 things.
it will either spread load across different cores but total combined cpu usage will not go above 1 core, so eg. on a dual core it will not use more than 50% of each core.
or it will only use 1 core.
Some high res flash videos used to saturate my core2duo, I would have frame stuttering whilst my 2nd core was idle and available but first core at 100%. Some web pages also capable of saturating cpu as content heavy again, not stressing entire cpu just one core.
So to be clear I mean one tab in a browser will only fully utilise one core at a time.
my i5 handles things much better as heavy pages cause it to trigger turbo mode.
Last edited by Chrysalis; 06-27-2010 at 11:14 PM.
task manager doesnt accurately measure utilization, especially with HT. a performance profiler would tell you whats truly going on during runtime. i.e. if an app has 4 threads 50% w/HT is the same performance as 100% w/o HT on my i7. i have observed and measured this before.
How do you measure "total throughput" ?
How does it relate to "integer throughput" and "floating point throughput", for which there are some generally accepted benchmark proxies?
Should we think of it as both int & fp independently? Their average? Sum? Minimum? Maximum?
Last edited by terrace215; 06-26-2010 at 03:46 PM.
operations per second, possibly in some abstraction.
depends on the benchmark. that's why you run multiple benchmarks.How does it relate to "integer throughput" and "floating point throughput", for which there are some generally accepted benchmark proxies?
no. they are just data types that should be completely transparent to the user. the kinds of benches you run should reflect what applications you use so you will already choose the best processor.Should we think of it as both int & fp independently? Their average? Sum? Minimum? Maximum?
you ask more questions than a 4 year old.
^
zzzz same old disc.
edit:
tz.. deleting a post...
Last edited by Hornet331; 06-27-2010 at 02:40 AM.
Posting Permissions
Reply With Quote
Bookmarks