8350 4.0ghz

>8350 4.0ghz
>ryzen 9 3.5ghz

I'm not even trolling, I'm just old as fuck. When I was young we only upgraded cpu to get more clockspeed. What is the point of upgrading if the clock speed is lower? Please be gentle, I don't understand.

Attached: anime-girl-sad.jpg (600x450, 25K)

I don't understand neither but look at benchmarks and see what procesor performs better

more cores, more cache

We used to have HDD the size of buildings and it would be 8MB big. Now we cab store 1 TB in SD card the size of finger nail.

Do you think we downgraded?

basically computer is told to add two numbers together, let's say
new cpus can do up 6 of these at the same time, so it will read ahead a bunch of instructions and do as many as it can in one clock cycle. bulldozer was a big retard and could only do like 2 of these at a time and couldn't read ahead as far

What's your point? My 2.0L 4-cyl car makes the same power than a 5.0L V8 from 25 years ago, whilst getting 42MPG on the highway. Shit advances, and efficiency gains come from a lot different little aspects of complex systems like CPUs or car engines.

It is so sad that you are an oldfag but still stupid

cache sizes, memory channels and instructions sets is what matters instead of mere clock speeds. nowadays it is also number of cores and how sufficiently they share the cache

today the rule of thumb is - more cache per core - the better. everything else is almost irrelevant

Shut the fuck up weeaboo and buy the cpu.

based retards

1 hz=one in out of the penis head. now you can have 5khz with a three incher or with an 8 incher. getmy drift, spaz?

Are instruction sets between AMD and Intel different of each other?

they are barely different externally. both execute machine code in the AMD64 instruction set with some extensions differing between Intel and AMD.

There are 3 things that are all almost equally important in regards to how strong a cpu is

IPC - How well the cpu performs at a certain clockspeed

Clockspeed - Self explanatory

Cores - Having more cores is like duplicating the CPU over and over again, it allows for more things to be done at the same time.

If a CPU is really good at two of these things but terrible at one, it wont perform well

High clockspeed, low ipc, high core count =bulldozer
Low clockspeed, high ipc, high core count = Slow
Low core count, high ipic, high clockspeed = Stutterfest

Thats basically how it all works

Attached: 1556123802995.png (150x262, 46K)

Zen1 at 3ghz is significantly faster than BD at 4ghz, my painfully retarded troll friend. The quadcore Ryzen APUs have better all around performance than the FX 8350.
Not even Excavator can hold up to low clocked Zen. Nothing in the Construction core family comes remotely close to performance per clock of Zen.

Clockspeed already reset once in 2005. If you're young enough to not remember that, you never actually lived through the era where every new CPU had significantly higher clocks.
Which means you're lying.

Attached: z94Of[1].png (1045x728, 66K)

this. sortof.

Cache is what matters most. Cache even arguably defines cores, as by looking at the cache on FX CPUs, you could actually consider them to have 2, 3, and 4 cores with SMT instead of 4, 6, and 8.

Cache does not define a core. You have no idea what SMT is.
The cores in the Bulldozer module are physical, separate execution paths, and are in no way similar to any implementation of SMT. They are literally the polar opposite.

Attached: 9.jpg (1732x1239, 323K)

>higher cache memory
>new ways of dealing with virtual memory
>new vector instructions
>more cores
all this matter for new processors to be faster, not clock speed

fucking this. Anyone remember Netburst? 15 years ago we were burning 4 GHz CPUs. And like 500 cycles per instruction, so it didn't matter for a shit, but yeah, we had 4 GHz CPUs in 2004. Clock speed means dick.

clock don't mean a fucking thing when comparing different architectures, even when both architectures are from the same company. I had an FX-9370 at 5.5GHz before. It got its ass handed to it by a stock i5-3570k in games and barely kept up with a stock i7-3770k in work loads. Then first gen Ryzen launched, a stock clocked R3-1200 quad core was giving me a run for my money.

By that definition Core 2 would be single and dual core processors as the cores on each die shared the massive L2 cache.

I define a core as having 1 set of front end and back end hardware, 1 full FPU and 1 full INT unit. By this definition, bulldozer would be considered "Core and a half" while architectures like Core and Ryzen would be considered full cores.

Thanks for the info. Sorry I left the house to go do some stuff.

>I left the house
It's like you're gay or something?

Even back in the Pentium 4 days clock speed was meaningless you dumb fuck because it didn't take into account instructions per clock.

So if I wanted to buy a ryzen cpu, I need an AM4+ motherboard, what do you suggest for a ryzen cpu?

Also, suck my fucking balls you gay homos.

plz no bully

Attached: 1464503186035.jpg (306x480, 24K)

>Not using a 2L v8 that revs to 14,200rpm
shiggy

WHAT RYZEN SHOULD I BUY

Why are modern 3cyl 1.0 turbo charged engines faster than ass old 4cyl 2.0 normal ones?

Wait for the 3700x, its coming out in less than a month

Many things. They can have better layout that improves them. Things like cache sizes. They can have wider instruction set support (though that's generally stable between modern cpus now). They can do better pipelining so each clock cycle does more at once.

And of course there's always more cores which is a nice multiplier on parallel workloads.

But the more the CPU can do faster the more memory it needs to be fed. It's already a major software optimization problem. Knowing how much typical software waits you'll probably find no substantial difference in applications which haven't been written for speed (exceptions: games, video editing, benchmarks). Seriously. When 99% of what your CPU is doing is waiting how fast it is doesn't matter much.
Upgrading your ram helps but it's not a place where we've seen much improvement.

which means OP is a troll then

Better IPC, so each clock can do more instructions. You can think of it like performance = IPC x clock speed. So if IPC is a lot higher and clock is slightly lower, performance is still better.

>what are cores
>what is IPC
Do you also think a single core of Zen 2 at 3.5 GHz is slower than a Pentium 4 EE at 3.7GHz?

Dumb underage anime poster.

Attached: 1542625521613.png (270x360, 66K)

Everyone has already made the argument about IPC, in processor caches, clock speed, and cores, but that isn't all there is to processor performance. Sometimes there's architecture related benefits, like supporting newer models of memory. Sometimes they implement some hardware baked algorithms for media encoding/decoding. There's a lot of nitty gritty details that barely get any mention on packaging or advertising because most buyers won't care. Also there's security related changes that go on behind the scenes which you will probably value.
>mfw I realize Pentium 4 is older than most posters here, and that's why nobody called this post out for assuming that's what OP was referencing when he said "when I was young".
It depends entirely on the features you need. If you're overclocking, you're going to avoid "A-series" mobos. If you're using multiple GPUs, a lot of hard drives, or whatever, you'll probably want an x-series motherboard.

Attached: deadinside.png (1280x720, 928K)

Not OP, but to be fair, in 2005 I was still in high school, and back then I didn't really care about technology aside from looking up porn and music on limewire.