AMD "Renoir" Ryzen 7 4700U Geekbenched

[btarunr]

An unknown Lenovo notebook powered by the 15-Watt Ryzen 7 4700U "Renoir" 8-core processor was put through Geekbench. The chip yielded scores of 4910 single-core, and 21693 multi-core. This puts the 4700U ahead of the Core i7-1065G7, which is known to score around 4400 points on average in the single-core test, and around 17000 on average, in the multi-core test, falling behind due to fewer CPU cores. The 4700U features an 8-core CPU based on the "Zen 2" microarchitecture. Its desktop compatriot, the Ryzen 7 3700X, is significantly faster, with around 20% higher single-core, and over 60% higher multi-core performance. This is probably because the 3700X is unconstrained with its 65-Watt TDP and significantly higher power limits. It also has four times more L3 cache, but that's probably to cushion the IFOP interconnect between the CPU chiplet and I/O die.



View at TechPowerUp Main Site

 

[lewis007]

Single core performance looks impressive.

 

[ratirt]

I'm seriously considering one of these in my next laptop. The numbers are pretty great.

 

[seronx]

*cough* 4700U */cough*

 

[wurschti]

*cough* 4700U */cough*

That looks to be correct. Also if I remember correctly 4700U has no SMT, meaning the 4800U will have some more performance gains.

 

[londiste]

Single-thread result lower than desktop is a bit strange but Zen2 does use 20W for 4.x GHz on one core, so 15W TDP might be why.

Can't wait for the reviews and especially the desktop versions of Renoir to come out.
Renoir should fix the problem with Ryzen 3000 high idle power consumption floor (about 10W) and being monolithic should have memory controller with lower latency.

 

[TheLostSwede]

Single-thread result lower than desktop is a bit strange but Zen2 does use 20W for 4.x GHz on one core, so 15W TDP might be why.

Can't wait for the reviews and especially the desktop versions of Renoir to come out.
Renoir should fix the problem with Ryzen 3000 high idle power consumption floor (about 10W) and being monolithic should have memory controller with lower latency.

Keep in mind that this is most likely an early system that might not have the final cooling solution or UEFI yet.

Looking forward to some proper reviews. My only real concern is battery life and if the notebook makers will finally put proper dual-channel memory support in their machines for AMD.

 

[notb]

Certainly an impressive score, but pointless comparison in general. This will always be heavily limited by the laptop. Ice Lake is only available in very slim models an. This score could be from a large gaming/workstation laptop. A comparison of mobile SKUs should always mention what laptops where used and in what conditions (battery or cable etc).

Also, I don't understand why both AMD and fans keep comparing to Ice Lake, when a more potent Comet Lake CPUs are available.

 

[ratirt]

Certainly an impressive score, but pointless comparison in general. This will always be heavily limited by the laptop. Ice Lake is only available in very slim models an. This score could be from a large gaming/workstation laptop. A comparison of mobile SKUs should always mention what laptops where used and in what conditions (battery or cable etc).

Also, I don't understand why both AMD and fans keep comparing to Ice Lake, when a more potent Comet Lake CPUs are available.

Gaming/workstation laptop? The specification for the processor is 15W and it stays like that. Shouldn't you focus on what the processor's specs are actually because this is a laptop area and the most important thing here is power consumption since that will determine where the laptop ends up.

 

[HwGeek]

I think @btarunr needs new glasses or another cup of Coffee, it's just 4700U LOL.

 

[ratirt]

I think @btarunr needs new glasses or another cup of Coffee, it's just 4700U LOL.

Yeah. I've noticed that too. He will adjust this eventually. Still 4700U is not just. Seeing the score it is really good. Considering it is a 15W CPU.

 

[TheLostSwede]

Yeah. I've noticed that too. He will adjust this eventually. Still 4700U is not just. Seeing the score it is really good. Considering it is a 15W CPU.

See slide above, OEM Configurable 12-25W and we have no idea about the configuration in this system.

 

[Melvis]

THat is some very impressive numbers from a 15/25w 8 core mobile CPU, these will be more then competative against intels current line up in the same 15/25w area. looks like I will be updating this year with numbers like that.

 

[HwGeek]

How this chip can go upto 4266Mhz? how will it compare against ryzen on desktop?
Also hope AMD will release Ryzen 9 4900H , unlocked with 3.5Ghz base, 4.5Ghz boost @ 54W TDP just to make lol against Intels 10980HK ;-).
this "4900H" + 4266Mhz mem could make really great "Laptop" ;-).

 

[ratirt]

See slide above, OEM Configurable 12-25W and we have no idea about the configuration in this system.

15-25W for this score is impressive. Configuration in the system you mean the frequency it will have or what will be the laptop it ends up in for?

How this chip can go upto 4266Mhz? how will it compare against ryzen on desktop?
Also hope AMD will release Ryzen 9 4900H , unlocked with 3.5Ghz base, 4.5Ghz boost @ 54W TDP just to make lol against Intels 10980HK ;-).
this "4900H" + 4266Mhz mem could make really great "Laptop" ;-).

Is AMD releasing a 4900U? I couldn't hear anything about it in the CES event and what the specs of it might be.
I don't think going 4266Mhz mem is a great idea considering the Infinity Fabric and its divider when mem frequency goes more than 3600Mhz?(was it 3600Mhz or more?) going more than that will actually make it slower.

 

[seronx]

How this chip can go upto 4266Mhz?

It uses a low-power double-data rate four x(LPDDR4X) interface for 4266 MHz.

I don't think going 4266Mhz mem is a great idea considering the Infinity Fabric and its divider when mem frequency goes more than 3600Mhz?(was it 3600Mhz or more?) going more than that will actually make it slower.

Fabric clock in Renoir is probably forced to asymmetrical 2:1, rather than 1:1.

.dram_speed_mts = 1600.0 => .fabricclk_mhz = 600.0 // DRAM @ 1600 MHz = IF @ 600 MHz
^-- probably a weird idle.
.dram_speed_mts = 2133.0 => .fabricclk_mhz = 1066.0 // DRAM @ 2133 MHz = IF @ 1066 MHz
.dram_speed_mts = 4266.0 => .fabricclk_mhz = 1600.0 // DRAM @ 4266 MHz = IF @ 1600 MHz
^-- If it was 3200 MHz, fabric clock would probably be also 1600 MHz.

 

[Valantar]

4700U is 8c8t.

 

[ratirt]

It uses a low-power double-data rate four x(LPDDR4X) interface for 4266 MHz.
Fabric clock in Renoir is probably forced to asymmetrical 2:1, rather than 1:1.

.dram_speed_mts = 1600.0 => .fabricclk_mhz = 600.0 // DRAM @ 1600 MHz = IF @ 600 MHz
^-- probably a weird idle.
.dram_speed_mts = 2133.0 => .fabricclk_mhz = 1066.0 // DRAM @ 2133 MHz = IF @ 1066 MHz
.dram_speed_mts = 4266.0 => .fabricclk_mhz = 1600.0 // DRAM @ 4266 MHz = IF @ 1600 MHz
^-- If it was 3200 MHz, fabric clock would probably be also 1600 MHz.

Come to think of it. The Renoir is a monolithic CPU. Is there an Infinity Fabric to communicate the core with iGPU?

 

[Valantar]

It uses a low-power double-data rate four x(LPDDR4X) interface for 4266 MHz.
Fabric clock in Renoir is probably forced to asymmetrical 2:1, rather than 1:1.

.dram_speed_mts = 1600.0 => .fabricclk_mhz = 600.0 // DRAM @ 1600 MHz = IF @ 600 MHz
^-- probably a weird idle.
.dram_speed_mts = 2133.0 => .fabricclk_mhz = 1066.0 // DRAM @ 2133 MHz = IF @ 1066 MHz
.dram_speed_mts = 4266.0 => .fabricclk_mhz = 1600.0 // DRAM @ 4266 MHz = IF @ 1600 MHz
^-- If it was 3200 MHz, fabric clock would probably be also 1600 MHz.

IF clock is fully independent from memory clock in Renoir. IIRC Anandtech reports this in the link in my previous post, and it was explicitly said by AMD at launch. No 1:1/2:1 ratios here.

Come to think of it. The Renoir is a monolithic CPU. Is there an Infinity Fabric to communicate the core with iGPU?

Yes. Otherwise the GPU would need to be connected by far less efficient and much lower bandwidth PCIe. Both IF and PCIe can be used on-die (APUs and Intel CPUs), inter-die (MCM packaging, see Zen2 and KBL-G for examples) and off-die. You need some sort of link between the CPU and iGPU, obviously, and AMD uses IF. On-die IF normally doesn't get much press though, as its normally not configurable or otherwise visible to end users.

 

[ratirt]

IF clock is fully independent from memory clock in Renoir. IIRC Anandtech reports this in the link in my previous post, and it was explicitly said by AMD at launch. No 1:1/2:1 ratios here.


Yes. Otherwise the GPU would need to be connected by far less efficient and much lower bandwidth PCIe. Both IF and PCIe can be used on-die (APUs and Intel CPUs), inter-die (MCM packaging, see Zen2 and KBL-G for examples) and off-die. You need some sort of link between the CPU and iGPU, obviously, and AMD uses IF. On-die IF normally doesn't get much press though, as its normally not configurable or otherwise visible to end users.

You need a link to get the iGPU and cores communicate, i thought maybe AMD didn't use IF. Anyway there's still correlation between the memory frequency and IF's speed. I don't think using 4266Mhz mem is a good thing for the CPU performance. I couldn't find some more specific stuff as for now. BTW you can have IF ticking at more than 1800Mhz. I thing the ceiling for the mem speed is 3733Mhz. that of course depends on the CPU but 1800Mhz IF is reachable especially if you use 3000 series processors. I'd assume for the 4000 series CPUS, it shouldn't be a problem achieving 1800Mhz IF.

 

[Valantar]

You need a link to get the iGPU and cores communicate, i thought maybe AMD didn't use IF. Anyway there's still correlation between the memory frequency and IF's speed. I don't think using 4266Mhz mem is a good thing for the CPU performance. I couldn't find some more specific stuff as for now. BTW you can have IF ticking at more than 1800Mhz. I thing the ceiling for the mem speed is 3733Mhz. that of course depends on the CPU but 1800Mhz IF is reachable especially if you use 3000 series processors. I'd assume for the 4000 series CPUS, it shouldn't be a problem achieving 1800Mhz IF.

No. AMD specifically designed Renoir to unlink DRAM and IF speeds entirely. They are separate clock domains with no relationship whatsoever in that design.

From AnandTech's launch article (there's a link here, experience tells me I have to make that explicit for you):

These CPUs all support LPDDR4X memory, up to 64 GB, and AMD says that the infinity fabric is not tied to this memory clock.

So what you're saying is true for desktop Zen2, but not mobile Zen2 (Renoir).


High speed IF can use quite a lot of power, so I would guess the mobile implementation is designed to hit a perf/w sweet spot but also possibly clock down when not needed. Unlinking this from memory clock speeds thus makes a lot of sense.

 

[ratirt]

No. AMD specifically designed Renoir to unlink DRAM and IF speeds entirely. They are separate clock domains with no relationship whatsoever in that design.

From AnandTech's launch article (there's a link here, experience tells me I have to make that explicit for you):

Nice. I didn't know that. Thanks.

 

[Chrispy_]

My biggest worry is that too much of the power and silicon budget has been moved away from the graphics cores and to the CPU cores. Both previous generations of APU have been perfectly adequate in the CPU department and sorely lacking in the IGP department, ranging from inadequate (Vega3 being as useless as low-end Intel HD) to acceptable-but-underwhelming - in that 720p30 might just be attainable in current titles.

In terms of CPU performance, 15W ultraportable customers really aren't clamouring for more cores. Those products are typically not multitasking mobile workstation powerhouses, lacking RAM, screen size, and storage for many serious workloads. To date, the vast majority of 3700U flagship APUs have been either 8GB or 16GB max, with 512GB NVMe 2x and 1080p displays. They are general-purpose consumption devices that could definitely use a little more GPU power but are typically beyond the point of diminishing returns in the CPU department already.

AMD keep saying that they have added 60% more performance per CU but then have stripped away 30-40% of the CUs The fact that AMD isn't singing praises about the 3D performance of its new 7nm APUs is a pretty bad sign of things to come, especially because the 4800U is likely to command a significant price premium. I would expect the 6CU option to sell at the same pricing tier as the previous 10CU 3700U models, which sucks because the 4800U will likely be price-competing against an MX250 or even faster dGPU options. I can only hope that LPDDR4X is used to its full advantage this generation. My 2700U shipped with a 13W TDP and single-channel DDR4 2400; The only reason I bought it was good cooling (for up to 25W) and an empty DIMM slot for dual-channel.

Is it too much to ask for a 15W ultraportable that provides a meaningful IGP upgrade? The bar is SO low, I'm honestly saddened by the lack of attempts to pass it.

 

[$ReaPeR$]

this looks very promising for the desktop versions of the 4k series ryzen. I cant wait for the benchmarks.

 

[londiste]

No. AMD specifically designed Renoir to unlink DRAM and IF speeds entirely. They are separate clock domains with no relationship whatsoever in that design.

Pretty sure that is not true. See @seronx post #16. The same "not tied" messaging has been used throughout AMD statements for Zen2. What AMD means by that phrasing is that IF frequency is not 1:1 to memory speed as it was for Zen/Zen+. They added a multiplier/divider and use that to facilitate faster memory speed.