• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Questions about transistors and clock speed

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Theocnoob

Member
Joined
Dec 1, 2007
Location
Near Toronto Canada
Hi,

Just curious about a few things:

Why are there 2x as many transistors on a first gen PIII as on a first gen P4 and yet the P4 family of processors is faster? Even taking into account the increased bandwidth and clock speeds?

Why, for example, does INTEL's 80core teraflop research chip (youtube it) only have 100 million transistors across 80 cores, but it can hit 1.6 teraflops, meanwhile my Core 2 has like 400 million transistors and it cant hold a candle to it?

Please enlighten me.
 
The first PIII was 28-100Mill+, P4 was 42-120Mil+. Cache sizes and number of internal pipes where the major differences.

The cell processors you are comparing have small caches and I think they run on native RISC code not requireing the X86 translators. Currently each core in the Xeon, Conroe and Kentsfields etc, contain large caches which use large numbers of tranys. The cell processors are also designed very stream lined which also cuts down on the trany count. It's an awsome processor, wish they would include them on the Fusion and Neha CPUs next year. That would make for major improvements if every thing from spreadsheets to gaming.
 
Its all about architecture. The design of the PIII chips is differant to that of a core2duo and very very different in the research chip (which probably couldn't even run a word processor). PIII's are less efficent by design, so they take more transistors to do the same amount of work as a p4, and the same again for a core 2. Also don't forget cache which shows up as a transistor count too. the more cache, the more transistors.

AlabamaCajun said:
The cell processors are also designed very stream lined which also cuts down on the trany count. It's an awsome processor, wish they would include them on the Fusion and Neha CPUs next year. That would make for major improvements if every thing from spreadsheets to gaming.
AlabamaCajun is referring to the Sony Cell processor i believe, an 8-core proc that is currently used in the PS3 (although only 7 cores are garrentied to work). They are very good.

So in conclusion, i wouldnt worry about it. it dosn't nessearily have any relation to the speed or capability of a processor core when comparing families.
 
the last P3's released where faster then some P4's at higher clocks. there is no way a P3 was less eff then a P4. in alot of cases a Tually celeron [email protected] with pc133 memory could match a P4 2.4ghz with DDR-333 memory. P3's had short pipes then the P4's, the now core 2 first got its go from Banis which is a highly modified P3 arch. then from there Dothan added double the L2 cache and so forth... even then when i was young 16 at the time i knew the P3 was the better cpu. even if it had pc133 memory, board that had RDRAM or DDR added for the P3. ever benifiteed from those memroy techs cause the cpu didnt need all that extra bandwidth they offered. this still holds true for core 2 cpus they dont need the max bandwidth at any ram bandwidth to be good. that has been one of the biggest thing for amd user is that the IMC could give the cpu more bandwidth to use. what good is that extra bandwidth if the cpu doesnt need it?
 
the last P3's released where faster then some P4's at higher clocks. there is no way a P3 was less eff then a P4. in alot of cases a Tually celeron [email protected] with pc133 memory could match a P4 2.4ghz with DDR-333 memory. P3's had short pipes then the P4's, the now core 2 first got its go from Banis which is a highly modified P3 arch. then from there Dothan added double the L2 cache and so forth... even then when i was young 16 at the time i knew the P3 was the better cpu. even if it had pc133 memory, board that had RDRAM or DDR added for the P3. ever benifiteed from those memroy techs cause the cpu didnt need all that extra bandwidth they offered. this still holds true for core 2 cpus they dont need the max bandwidth at any ram bandwidth to be good. that has been one of the biggest thing for amd user is that the IMC could give the cpu more bandwidth to use. what good is that extra bandwidth if the cpu doesnt need it?

He's speaking of the pre-Tually parts (sorry forget what the cores were called not the 733,800,833 and 933s etc. Intel really was going in the right direction with Tually but they thought industry could benefit form netburst. Hyperthreading did help a lot with day to day tasks because it worked well with mixed apps where overlapping IO and Ram use requests allowed for the channeling. Games on the other hand being monolithic at the time suffered compared to the Athlons that were becoming popular.
 
ah ok.. well AC the per-tuallies the ones you speak of are the coppermine core. still solid cpus for there day, the P3 700-E @933 was one of the most common combo's for ocing. heheh my friend next door took that box from me about 6months ago its still running just fine to this day...
 
Back