This is mean't to test what you know about your hardware. Even if you do not know, POST YOUR GUESS!!!
Printable View
This is mean't to test what you know about your hardware. Even if you do not know, POST YOUR GUESS!!!
32 bit fo' sho' m0f0
I am surprized! Many, who I know personally, thought that the Pentium was a 64 - bit CPU. Intel actually put 2 486DX's on the same die and added enhancements. But this is the pipeline. The chip itself as a whole is only 32 - bits...
Don't make me he-***** man-slap them!
Do it... They made us think of the Pentium as a 64 - bit CPU... It's sad to think of a gaming console being more powerful than a PC:
N64 = SGI MIPS R4000 @ 93.5 MHz - 1994( for the CPU)
Equivalent to a 187 MHz 32 - Bit Pentium CPU - ~1997
Hmmm... factor in that SGI is known for OUTSTANDING graphics and Floating Point Ops in it's CPUs...
I think this'll turn into a debate! :D
Well Intel chips suck....so there. They just soop up the clock speeds...thats about it.
btw - wasnt Intel's 64-bit chip (AI-64 or something) due out last year?
The Itanium is already out, as is a much-unspoken-about 64-bit Windows.
Nobody ever pretended the Pentium was 64bit ....
And yeah the original P5 was basically 2 486s, but its moved on quite a bit from there.
Isn't the DEC Alpha have a 64bit chip? I'm sure there were 64bit RISC processors long before Intel got around to it.
Yeah 64bit cpus have been out for donkeys years at this stage.
But its only when wintel comes out with something that everything pays attention
Yeah, the Pentium 4s are just 4 486es running at fast clock speeds :rolleyes: :p :pQuote:
Originally posted by plenderj
Nobody ever pretended the Pentium was 64bit ....
And yeah the original P5 was basically 2 486s, but its moved on quite a bit from there.
Speaking of 64-bit, the upcoming AMD Clawhammer = teh winnar. It isn't as fast with 64 bit stuff..but is MUCH cheaper and executes 32 bit apps much better than Itanium, and even better than any 32 bit CPU out today.
They put out a winnar with Athlon, made a cheap winnar with Duron, so they can do it again with Clawhammer...w00t :D
'Course, every generation of Motorola's have outperformed Intels, but no one cares.Quote:
Originally posted by plenderj
Yeah 64bit cpus have been out for donkeys years at this stage.
But its only when wintel comes out with something that everything pays attention
Also in sad news, Compaq bought DEC and then sold off the Alpha processor to Intel. I think Compaq ported their Tru64 OS first, but Intel is just going to shove Alpha in a drawer.
If you can't beat 'em, bury 'em.
yeah, as far as IPCs go, Motorola CPUs = winnar.Quote:
Originally posted by CiberTHuG
'Course, every generation of Motorola's have outperformed Intels, but no one cares.
Also in sad news, Compaq bought DEC and then sold off the Alpha processor to Intel. I think Compaq ported their Tru64 OS first, but Intel is just going to shove Alpha in a drawer.
If you can't beat 'em, bury 'em.
AMD CPUs a distant second.
Yes, the DEC Alpha was a 64-bit RISC CPU. I say 'was' as DEC has gone under(I believe). A pity, really. Their Alpha 21264(as it was called) was running at 600MHz @ 64 - bit when Intel was stuck with the 150MHz Pentium... The Alpha 21264 actually held the record in the Guiness book of records for several years(at least 2) for the fastest chip. Check this: I was looking at an OLD computer magazine from about 1995 and there was a 200MHz DEC Alpha... The same magazine showed the 60MHz Pentium as "top-of-the-line".
I just realized your last response, Ciber THuG, and it seems I am confirmed. DEC is gone... In the mess of one of Intel's drawers... GL finding the alpha again... :(
Oh well... The dual 486's on a single die explains why the Pentium is 2 times faster than a 486 at the same speed. Another thing: The 486SX was a mistake: Here this: The 486SX was EXACTLY the same as the 486DX except for the fact the the FPU was DISABLED. CHEAP B'ARDS :mad: I hear it was an accident - a design flaw. But intel sells it cheaper. The 487 is a 486DX(With FPU ENABLED). That's it! I think it was just a scheme the make us spend more money... :mad: The 487 basically cuts the 486SX out and replaces it with itself. Try taking out the 486SX(If you got one).
There is a fault here. Sorry nishantp, but Intel doesn't just soup up the clock speeds. They have to rewire some of the chip to make the pathways shorter. Let me explain:
If a CPU runs at 33 MHz, then one clock cycle(One instruction) MUST be executed in 30ns(Nanoseconds - a billionth of a second). Soup up the clock speed to 100MHz... Now each instruction must be executed in 10ns. If the clock pulse doesn't execute in that time, the chip creates faults as the results of one instruction run into results of another. As far as overclocking goes, you can only overclock so much or the CPU stops functioning. The higher the clock speed, the more things that happen in a given time, the hotter the chip gets. If you go to 1000MHz or 1GHz, each instruction must be executed in 1ns. Three options: rewire it, move the die to a smaller process(maybe .015micron), OR move the chip design to a copper die.
AMD is finally catching up to Intel. At the AMD K6, the Intel chip of same speed out performed the K6. Introduce Athlon. A 1.7GHz Athlon is about the same as a Pentium 4 at 1.4GHz. Believe it or not! And the Athlon's also cheaper.
Speaking of a CPU's data bus width... Why don't our PC's move to 128 - bit? The Gamecube, Dreamcast, and PS2 are all 128 - bit... And then there's sad X-Box using Intel P3 @ 733MHz(Which crashed when one of my friend tested it - heh an M$ product). Think of a CPU with a wider bandwidth as also a faster CPU - It can process more at once.
One last laugh from me and then I'm done with my speech - for now :D. Did you know that Intel does not use their own chips in the design process? I've heard that they use Apple Mac's with the Motorola 68K CPU!!! Isn't that funny!? Maybe they don't want the FDIV bug(Pentium Classic - 60/66MHz - Divide error) to be passed on... Or maybe they need a more stable system :eek:
lol :D
As far as wider bandwidths go, have you heard of a vector processor? :)
We're talking Crays here ;)
You mean the other way around..1.4ghz athlon outperforms a 1.7ghz p4 ;)Quote:
Originally posted by Warmaster199
AMD K6, the Intel chip of same speed out performed the K6. Introduce Athlon. A 1.7GHz Athlon is about the same as a Pentium 4 at 1.4GHz. Believe it or not! And the Athlon's also cheaper.
*thinks of a 1.7ghz Athlon vs a 1.4ghz p4...mwahahaha*
and the NEW Palomino core, a true 1.4ghz Palomino (Athlon XP) will outperform a P4 2.0.
I am sorry... You are right... I honestly mean't 1.7GHz P4 = 1.4GHz Athlon...
Never heard of the Palomino Core... Is it really that fast???
Oh... I forgot something... Isn't it odd that Intel chips are several million transistors and then there's AMD... a couple 100,000 less(Or couple million)... No wonder AMD is faster. The chips are smaller. Smaller is faster
Palomino = Athlon XPQuote:
Originally posted by Warmaster199
I am sorry... You are right... I honestly mean't 1.7GHz P4 = 1.4GHz Athlon...
Never heard of the Palomino Core... Is it really that fast???
The AMD Athlon has approximately 37 million transistors.
I am waiting for the AMD Athlon XP 2GHz to be mid range so I can get a DDR motherboard... mmmmm DDR RAM..... Gonna have to get a GeForce 3 as well..... *drools*
Whoa... I'm pretty sure that the Pentium 4 has more transistors. Doesn't the P4 have a larger instruction set as well??? The SIMD instructions as well as SSE(Or whatever), and MMX, and the 80x86 instructions... The AMD Athlon just has the 80x86 instructions, and the 3DNOW technology, right? Or is there more(Of course these expand in every new chip design)...Quote:
Originally posted by plenderj
The AMD Athlon has approximately 37 million transistors.
Isn't DDR RAM the same as Rambus RAM(RDRAM)??? It's speedy like crazy: 2 writes/read per clk pulse, with 2 banks, you can access two banks simultaneously. I seen on an Intel Desmonstration that RDRAM does 3.6GBytes/sec max transfer rate!!! Why you would need that kind of speed I do NOT know... But it is good. Video RAM is about that speed as well...
GeForce3? They're coming out with a GeForce3????? OMG! Imagine how awesome that would be. GeForce2 already rocks!(ATI Radeon is better than GF2). The only competitor for nVidia is ATI. Remember: New video cards come out every 6 months(I read it in a magazine).
Jamagei, don't forget the 100GByte IBM Deskstar Hard Disk... 100MBytes/sec transfer rate will satisfy the need for speed(for an HDD)!
We shouldn't need all these extra instructions.
My opinion is pretty much echoed by Sun's documentation for the SPARC - RISC is better.
And even Intel has (internally) admitted to that with a lot of their new stuff.
That just screams "sarcasm" :DQuote:
Originally posted by Warmaster199
Whoa... I'm pretty sure that the Pentium 4 has more transistors. Doesn't the P4 have a larger instruction set as well??? The SIMD instructions as well as SSE(Or whatever), and MMX, and the 80x86 instructions... The AMD Athlon just has the 80x86 instructions, and the 3DNOW technology, right? Or is there more(Of course these expand in every new chip design)...
Isn't DDR RAM the same as Rambus RAM(RDRAM)??? It's speedy like crazy: 2 writes/read per clk pulse, with 2 banks, you can access two banks simultaneously. I seen on an Intel Desmonstration that RDRAM does 3.6GBytes/sec max transfer rate!!! Why you would need that kind of speed I do NOT know... But it is good. Video RAM is about that speed as well...
GeForce3? They're coming out with a GeForce3????? OMG! Imagine how awesome that would be. GeForce2 already rocks!(ATI Radeon is better than GF2). The only competitor for nVidia is ATI. Remember: New video cards come out every 6 months(I read it in a magazine).
Jamagei, don't forget the 100GByte IBM Deskstar Hard Disk... 100MBytes/sec transfer rate will satisfy the need for speed(for an HDD)!
Although I am an ATI fan--go ATI for making fast, low priced video chips bundled with the All-In-Wonder.
True. But the object of this CISC is to have many instructions to make programming them easier and quicker. Less instuctions are needed in your program's code, because a single complex instruction on an Intel or AMD chip can be the same as 5, 10, even 100 instructions called on a RISC based chip. This leads to a CISC machine needing less RAM than a RISC machine.
BUT: RISC is better. RISC means the instruction set IS smaller, meaning the chip supports less instructions. These are basic bare instructions, BUT they can be executed FASTER because the FSM(Finite State Machine - Control unit) needs to decode less and there is less overhead(Less instructions to go through). This means that the die is smaller(faster), there are less transistors(faster, AND cooler running!!! YAY! SMALLER HEATSINK!), and that you MAY need to execute more instructions to do a certain thing.
Overclockers probably love RISC CPUs because they run cooler(as I said). This means that they could be overclocked higher that CISC machines because there is less transistors - less heat, and you can simply put a bigger heatsink on it... OR a heat sink with a fan on it!
One last thing on RISC. I read this on the net: A RISC CPU can works in layers... Let me explain:
- Get Operation
- Decode it
- Execute it
- Output it
With a CISC machine, it waits for a single instruction to go through the whole thing then does the next operation. If I understand correctly, then RISC CPUs have an instruction running in "get operation" while another instruction is in "Decode" while another is in "Execute" and another is in "Output". This would lead to 4 INSTRUCTIONS PER CLOCK CYCLE!!! So if you got a 1 GHz RISC CPU it's 4 billion inst/sec while a CISC would be 1 billion ops/sec.
Am I(Or the article) right!?
Well, sort of. RISC processors do that but so do most modern CISC processors. There are often multiple instructions at different stages in the processor, although not necessarily one instruction per stage.
On the subject of RISC vs CISC, it's interesting to note that both the Athlon and (I believe) the later Intel chips are fundamentally RISC with an extra abstraction layer of microcode embedded into them, to implement a CISC instruction set.
Well, the M68K processor is (I believe) a RISC processor.
And I was studying ASM for the M68K last year, and I quite liked it. Took a while to get my head around branch instructions, and how they're used to impliment loops or if statements and what not, but at the end of the day, 'twas a beut of a cpu !
Athlon XP now has added 54 new SSE compatible instructionsQuote:
Originally posted by Warmaster199
Whoa... I'm pretty sure that the Pentium 4 has more transistors. Doesn't the P4 have a larger instruction set as well??? The SIMD instructions as well as SSE(Or whatever), and MMX, and the 80x86 instructions... The AMD Athlon just has the 80x86 instructions, and the 3DNOW technology, right? Or is there more(Of course these expand in every new chip design)...
Rambus = 800MHzQuote:
Isn't DDR RAM the same as Rambus RAM(RDRAM)??? It's speedy like crazy: 2 writes/read per clk pulse, with 2 banks, you can access two banks simultaneously. I seen on an Intel Desmonstration that RDRAM does 3.6GBytes/sec max transfer rate!!! Why you would need that kind of speed I do NOT know... But it is good. Video RAM is about that speed as well...
DDR = 2100MHz
Yeah but whats the point in having such extremely fast ram, if you're still stuck with a relatively slow system bus ?
Quote:
Originally posted by plenderj
Yeah but whats the point in having such extremely fast ram, if you're still stuck with a relatively slow system bus ?
Well, because AMD doesnt use Rambus Ram, and if you want a better processor, then you have to either use DDR or SDRam, and DDR is definitely faster. Anyone whos interested i have a benchmark between P4s with Rambus and Athlon XPs with DDR. And several others... I havent sold a system yet with an intel processor, though ive been asked for them almost every time. Look at the performance, and youll see.
If you put a P4 at the same clock speed at a P3, you will get better performance (20% better) out of the P3. Intel is playing a MHz game that is only designed to confuse consumers.
P4 have increased clock speed so rapidly only by lowering the Instructions per Clock Cycle (IPC). It says right in Intel's Literature about their P4. "Performance = MHz X IPC". Why then did they lower one and raise the other?? Because 2GHz sounds really cool to consumers!! Not to mention that the P4 is about double the price, for the same performance.
And then theres the Thermal Diode (which both processors have). The only difference is that AMD has a thermal diode that is controllable by a compatible BIOS, and Intel has one that just goes ahead and slows down your processor for you when it hits what it considers to be a "Dangerous" temperature. (95°F) This can cause a serious decrease in performance, but hey, if you're going to toss your computer around enough that your cooling fan will fall off, then at least your processor wont fry huh??