News AMD Ryzen 3 Meets Zen 2: New Quad-Core CPUs Would Rival Intel Core i3 Comet Lake

[Admin]

AMD is reportedly working on Ryzen 3 3300X and Ryzen 3 3100 CPUs that would compete with Intel Comet Lake.

AMD Ryzen 3 Meets Zen 2: New Quad-Core CPUs Would Rival Intel Core i3 Comet Lake : Read more

 

[pug_s]

I doubt that these cpus would sell because it doesn't have an integrated gpu. If they want 4 cores 8 threads they can get the 3400g

 

[King_V]

Maybe - but do the majority of i3 CPUs (not business-use) wind up using the integrated graphics? Do they sell F variants of the i3?

 

[TCA_ChinChin]

I don't understand why AMD is expected to lower TDp for lower core count parts when the same is not expected from Intel. Their core i3 has the same TDP rating as their i7 and i9 processors so why is it not expected for those i3 processors to have reduced TDP?

 

[TJ Hooker]

I doubt that these cpus would sell because it doesn't have an integrated gpu. If they want 4 cores 8 threads they can get the 3400g

The difference there is that the current 3000 series APUs are Zen+, while these new Ryzen 3s would be Zen 2.

Maybe - but do the majority of i3 CPUs (not business-use) wind up using the integrated graphics? Do they sell F variants of the i3?

They sure do! The 9100F seems to be significantly cheaper than other i3s, and is apparently the best selling i3 on Amazon.

 

[watzupken]

I doubt that these cpus would sell because it doesn't have an integrated gpu. If they want 4 cores 8 threads they can get the 3400g

I agree at the low end, it makes sense to integrate a GPU on the chip. So instead of creating a separate line of Ryzen 3, AMD should just utilize their APUs to fill in this gap.

 

[watzupken]

"The most interesting part about the two unannounced Ryzen 3 chips is that they reportedly adhere to a 65W TDP (thermal design power), which is the same for 2000-series Ryzen 3 chips and 3000-series Ryzen 5 parts. With the Zen 2 optimizations and the die-shrink to the 7nm process node, one would expect the quad-core Zen 2 parts to come with a lower TDP. "

I see no point in picking on the TDP here. I think its clear for sometime that the TDP is not an accurate way to measure the actual power requirements. How sure are we that Intel is running within the 65W in the first place, and given whatever boost rate they allow on these low end chips, it is just going to blow past the rated TDP easily.

 

[alextheblue]

I agree at the low end, it makes sense to integrate a GPU on the chip. So instead of creating a separate line of Ryzen 3, AMD should just utilize their APUs to fill in this gap.

They could very well do both, and may even use the same die, as they often do.

 

[cryoburner]

There are two ways that AMD could spit out a quad-core Ryzen 3 chip. It could salvage imperfect Ryzen 5 3600 chips and disable the defective CCX or utilize an entirely new design, where there's only a single CCX inside the CCD. The latter would help AMD become more cost-competitive against Intel's line of Core i3 CPUs.

The first option could also be cost-competitive, perhaps even more so considering the investment that would likely be required to design an all-new single-CCX chip for the budget market. I imagine AMD has likely built up a collection of Zen 2 chiplets over the past year that have an entire CCX not meeting the standards for their other chips (which don't disable more than 2 cores per 8-core chiplet), so they might as well make use of them. I kind of doubt that AMD would make a brand new chip design for this market when everything from a $175 Ryzen 3600 up through their multi-thousand dollar Threadripper and Epyc CPUs are built using the same chiplets. If they did, it would be in the form of an APU with integrated graphics, and would probably be branded as a 4000-series chip.

 

[TerryLaze]

I don't understand why AMD is expected to lower TDp for lower core count parts when the same is not expected from Intel. Their core i3 has the same TDP rating as their i7 and i9 processors so why is it not expected for those i3 processors to have reduced TDP?

14nm + iGPU against 7nm without iGPU you would expect one of those two to have a considerably lower TDP just due to that.

Their core i3 has the same TDP rating as their i7 and i9 processors so why is it not expected for those i3 processors to have reduced TDP?

The i3s do have a lower TDP.

 

[TerryLaze]

"The most interesting part about the two unannounced Ryzen 3 chips is that they reportedly adhere to a 65W TDP (thermal design power), which is the same for 2000-series Ryzen 3 chips and 3000-series Ryzen 5 parts. With the Zen 2 optimizations and the die-shrink to the 7nm process node, one would expect the quad-core Zen 2 parts to come with a lower TDP. "

I see no point in picking on the TDP here. I think its clear for sometime that the TDP is not an accurate way to measure the actual power requirements. How sure are we that Intel is running within the 65W in the first place, and given whatever boost rate they allow on these low end chips, it is just going to blow past the rated TDP easily.

TDP is how much heat a CPU produces,and not even that it's how much of that heat you need to get away from the CPU to keep it below the cut off temp (105° for intel) ,it's not how much power the CPUs draw.

 

[TJ Hooker]

The i3s do have a lower TDP.

Not for the last couple generations. E.g. an i3 9100 has the same TDP as an I9 9900.

TDP is how much heat a CPU produces,and not even that it's how much of that heat you need to get away from the CPU to keep it below the cut off temp (105° for intel) ,it's not how much power the CPUs draw.

CPU heat generation is essentially equal to power draw. And under steady state conditions, i.e. where temp is not increasing, required heat dissipation is equal to CPU heat generation.

In theory TDP should be approximately equal to power draw under a sustained, all core, heavy-ish load. But that seems to be becoming less and less true in recent generations, for both Intel and AMD.

In Intel's case it's because mobo manufacturers are tending to set turbo boost power and duration to be effectively unlimited by default, such that TDP becomes meaningless.

 

[TerryLaze]

CPU heat generation is essentially equal to power draw. And under steady state conditions, i.e. where temp is not increasing, required heat dissipation is equal to CPU heat generation.

So you think that any CPU has to be running at 0°C?

 

[King_V]

I wouldn't worry too much about the TDP rating. The Athlon 200GE, 220GE, 240GE, and 3000G are all rated at 35W, yet, in the order I listed, they have successively higher clocks speeds. There's no way that they're all producing the same amount of heat at stock speeds.

They probably just did the same thing here. I would think if any were guaranteed to stay below a lower number, they'd be touting that, but, for whatever reason, they didn't.

 

[cryoburner]

I wouldn't worry too much about the TDP rating. The Athlon 200GE, 220GE, 240GE, and 3000G are all rated at 35W, yet, in the order I listed, they have successively higher clocks speeds. There's no way that they're all producing the same amount of heat at stock speeds.

Yeah, TDPs tend to often get rounded up to certain fixed numbers. I think AMD is pretty much just saying that they recommend a cooler with at least that level of cooling capacity, not necessarily that the processor will be using that much power under typical use. As far as stock coolers for their desktop processors go, AMD has generally been pairing the "65 watt" models with the Wraith Stealth, the "95 watt" models with the Wraith Spire, and the "105 watt" models with the Wraith Prism, though there are some exceptions where they provide a higher-end cooler than the TDP would deem necessary, like with the 3700X. Processors that they put into the "35 watt" bracket, namely the Athlons, just get a small rectangular no-name cooler.

So about the only thing to assume based on the "65 watt" TDPs of the 3300X and 3100 is that they will probably ship with the Wraith Stealth cooler. Actual power use and heat output could be a fair amount lower than their predecessors though. Of course, if they are using the same design as the 6 and 8-core parts, it's possible that heat output might not be much better than previous Ryzen 3s, since they would still be using a 12nm IO chip and have several disabled cores.

 

[TJ Hooker]

So you think that any CPU has to be running at 0°C?

What? The particular temperature the CPUs are running at is irrelevant to the statement of mine that you quoted.

 

[TerryLaze]

What? The particular temperature the CPUs are running at is irrelevant to the statement of mine that you quoted.

The thing is that you are talking about steady state so your argument is that the CPU temp should always stay the same as it is when the CPU is powered on (0 degrees or whatever room temperature is) but that is not how computers and TDP works.
TDP does allow for temperature to increase,it even allows for clocks to be reduced within stated stats.

"And under steady state conditions, i.e. where temp is not increasing, required heat dissipation is equal to CPU heat generation. "

 

[TJ Hooker]

The thing is that you are talking about steady state so your argument is that the CPU temp should always stay the same as it is when the CPU is powered on (0 degrees or whatever room temperature is) but that is not how computers and TDP works.
TDP does allow for temperature to increase,it even allows for clocks to be reduced within stated stats.

"And under steady state conditions, i.e. where temp is not increasing, required heat dissipation is equal to CPU heat generation. "

No, steady state does not mean that the values cannot change from what they were at time = 0 (i.e. when you turned the PC on). It means the values have settled at some final value once some steady load has been applied for some time.

E.g. you start running Prime95. Your CPU temp starts rising rapidly but eventually levels out after you've run it for a while. You are now at steady state. At this point the heat being dissipated is equal to power draw. In other words, if your temp is not changing then heat dissipation must equal power draw. If they weren't equal, then temp would keep increasing indefinitely. Or in this case, until your CPU overheated and throttled, thereby reducing power draw such that it's lower than or equal to dissipation.

Edit: Although for a PC (and most real world examples) your ability to dissipate power will increase as the device temperature goes up. Because heat flow is proportional to the temperature difference between the CPU/GPU/etc. and ambient temp. So if you aren't able to dissipate all the power the CPU is drawing at temperature X and therefore the temp starts increasing, you may still stabilize at some higher temperature. You just have to hope you hit the temperature where heat dissipation matches power draw before your CPU starts overheating

 

[TerryLaze]

No, steady state does not mean that the values cannot change from what they were at time = 0 (i.e. when you turned the PC on). It means the values have settled at some final value once some steady load has been applied for some time.

E.g. you start running Prime95. Your CPU temp starts rising rapidly but eventually levels out after you've run it for a while. You are now at steady state. At this point the heat being dissipated is equal to power draw. In other words, if your temp is not changing then heat dissipation must equal power draw. If they weren't equal, then temp would keep increasing indefinitely. Or in this case, until your CPU overheated and throttled, thereby reducing power draw such that it's lower than or equal to dissipation.

Yes absolutely correct and with the TDP(watt cooling) intel states their CPUs will stay at or below 100-105 degrees at steady state and will run at base or between base and all core turbo at steady state, base clocks for more demanding things turbo for less demanding things, i.e. default settings.

You will never get ~250W power draw out of an 9900k with a normal cooler.

 

[TJ Hooker]

Yes absolutely correct and with the TDP(watt cooling) intel states their CPUs will stay at or below 100-105 degrees at steady state and will run at base or between base and all core turbo at steady state, base clocks for more demanding things turbo for less demanding things, i.e. default settings.

You will never get ~250W power draw out of an 9900k with a normal cooler.

I agree, a 9900K should be able to run any load indefinitely at least at its base clocks with power draw =< TDP (i.e. 95W). But I don't think most people who bought a 9900K are going to be happy with it potentially running 1 GHz or more below max rated boost clocks in certain applications.

 

[TerryLaze]

But I don't think most people who bought a 9900K are going to be happy with it potentially running 1 GHz or more below max rated boost clocks in certain applications.

Most people go for much better cooling and for overclocking,they don't expect the CPU to run at minimums,that's why they buy the k version. (and why intel doesn't actually pack in a 95W cooler with them)