AMD's Radeon 780M integrated graphics hits 3.15GHz in Ryzen 7 8700G overclocking test — 37% faster on average with iGPU and memory overclock
15% faster with automatic tools
Overclocking the Radeon 780M integrated GPU in the Ryzen 7 8700G and its DDR5 memory could yield up to 37% better performance on average per benchmarks from overclocker SkatterBencher. The extreme overclocker, who has several world records under his belt, was able to overclock the 780M iGPU to 3.15GHz and DDR5 RAM to 8,000MHz to achieve an average 37% boost across all kinds of workloads. Even an overclock on just the 780M could see a 22% increase in performance.
In his detailed analysis, SkatterBencher investigated five methods for improving the 8700G's gaming performance, starting with the low-hanging fruit of enabling PBO and EXPO, then manual overclocking on the iGPU, CPU, and memory. The 8700G and its 780M graphics were run through tests like Geekbench 6 GPU and 3DMark Night Raid, and three games at 1080p, including Returnal.
Row 0 - Cell 0 | Average Performance | iGPU Clock Speed | iGPU Power in Speed Way |
Stock | 100% | 2,881MHz | 50W |
PBO and EXPO | 115.68% | 2,903MHz | 69W |
Tuned PBO and EXPO | 119.2% | 3,099MHz | 83W |
Manual iGPU OC and EXPO | 122.31% | 3,086MHz | 156W |
Manual iGPU and CPU OC and EXPO | 122.69% | ~3,086MHz | ~156W |
Manual iGPU, CPU, and RAM OC | 137.75% | N/A | N/A |
With completely manual overclocking, SkatterBencher was able to achieve a peak clock speed of 3,150MHz or 3.15GHz on the 780M, and a sustained frequency of 3,086MHz. That's a fair bit higher than the stock clock speed of 2,900MHz (or 2,881MHz according to the actual benchmark numbers).
The fully manual CPU, iGPU, and memory overclock resulted in a 37.75% performance boost over the default 8700G, pushing the chip very close to low-end desktop graphics card territory. The overclocked APU's score of 39,427 sidles right up to the RX 6400's 41,211 points in a quick test for comparison. There is, though, a bit of an asterisk on this score, as SkatterBencher says the memory overclock wasn't entirely stable. Perhaps a slightly lower memory clock speed with an improvement in the mid-30s percent is more realistic.
Users who don't want to overclock and find the perfect frequency and timings for their RAM are in luck, however. Merely enabling PBO and EXPO achieved 15% higher performance, and that only required enabling those settings in the BIOS. A 22% performance boost was achieved by upgrading from PBO to a manual GPU frequency target of 3,150MHz at 1.2 volts, which is much simpler than manual tuning PBO.
Though it's clear that overclocking the 8700G can transform it from a good APU into a chip that's about on par with an actual graphics card, it's still clear it can't really replace low-end GPUs like the RX 6400, RTX 3050 6GB, and Arc A380. The 8700G is still really expensive for its performance level, costing $330 on its own when one of these low-end desktop cards run for $150 at most and can be paired with a cheap $100 to $150 CPU. The 8700G will also require high-end DDR5 memory if you want to have any chance of achieving an 8,000MHz frequency on the RAM, which clearly influences the 780M's performance the most.
Plus, the 780M gets very hot with an overclock, and consumed 156 watts with a manual GPU overclock. That's more power than the RTX 4060 uses, which is absurd considering how much faster the 4060 is. Even with PBO (tuned or otherwise), the 780M eats up 69 to 83 watts, which is equivalent to the 6400 and its 75 watt TDP. Under a combined CPU and GPU workload, SkatterBencher showed the 8700G could hit a whopping 260 watts with manual overclocking. In other words, you're going to need a very good cooler for a heavily overclocked 8700G.
Stay On the Cutting Edge: Get the Tom's Hardware Newsletter
Get Tom's Hardware's best news and in-depth reviews, straight to your inbox.
Matthew Connatser is a freelancing writer for Tom's Hardware US. He writes articles about CPUs, GPUs, SSDs, and computers in general.
-
Bluoper Bit of a goofy ahh chip.Reply
On a more serious note, how do you think the 780m would perform if it had a large amount of 3d v-cache? -
jeremyj_83
If it is used like the Infinity Cache on the GPUs it could effectively double the bandwidth from the RAM. That would really help the GPU performance.Bluoper said:Bit of a goofy ahh chip.
On a more serious note, how do you think the 780m would perform if it had a large amount of 3d v-cache?
"Thanks to the 128MB cache, the framebuffer mostly ends up being cached, which drastically cuts down memory access. AMD says the effective bandwidth of the GDDR6 memory ends up being 119 percent higher than what the raw bandwidth would suggest."
https://www.tomshardware.com/news/the-amd-radeon-rx-6800-xt-and-rx-6800-review -
usertests
What jeremy said, except technically 3D V-Cache is inaccessible to iGPU, just like any of AMD's CPU L3 caches. There was that moment where some website reported quadrupled iGPU performance with the 7800X3D but it was just a mistake.Bluoper said:Bit of a goofy ahh chip.
On a more serious note, how do you think the 780m would perform if it had a large amount of 3d v-cache?
So what we'd want is Infinity Cache to come to mainstream APUs, or even 3D Infinity Cache, or AMD changing how it handles L3 to allow it to be shared between CPU + GPU. No evidence seen that they will do any of that for Strix Point. Not sure about Halo.
Even a measly 16 MiB might be enough to get the ball rolling. That's the amount in Navi 24 (6500 XT and 6400). -
jeremyj_83
I'd want the 128MB like what the i7-5775c had for eDRAM. However, that would make the chip a lot more expensive.usertests said:What jeremy said, except technically 3D V-Cache is inaccessible to iGPU, just like any of AMD's CPU L3 caches. There was that moment where some website reported quadrupled iGPU performance with the 7800X3D but it was just a mistake.
So what we'd want is Infinity Cache to come to mainstream APUs, or even 3D Infinity Cache, or AMD changing how it handles L3 to allow it to be shared between CPU + GPU. No evidence seen that they will do any of that for Strix Point. Not sure about Halo.
Even a measly 16 MiB might be enough to get the ball rolling. That's the amount in Navi 24 (6500 XT and 6400). -
hotaru251 so 20% improved performance at the cost of 3x the power consumption & thermals are crazy?Reply -
King_V
Yeah, that's a mess. I mean, this performance uplift is pretty cool in that it could be accomplished, in the sort of "CAN it be done?" kind of way, but it's not really practical, and absolutely not worth the power-use and thermal management headaches that are the price of that performance gain.hotaru251 said:so 20% improved performance at the cost of 3x the power consumption & thermals are crazy?
But, of course, the chip wasn't really designed to be pushed like this. -
usertests
Intel's Adamantine would have brought that back (L4 cache, usable by iGPU) as soon as Meteor Lake. But it has been a no-show so we'll have to wait for Arrow Lake at least. Size could be similar too, I heard 128/512 MiB.jeremyj_83 said:I'd want the 128MB like what the i7-5775c had for eDRAM. However, that would make the chip a lot more expensive. -
greenreaper Note for anyone trying this kind of thing: check that your GFX Curve Optimiser settings are actually set. My experience on the B650E-I is that ASUS's control for GFX CO doesn't work, as shown by a lack of voltage changes - you have to set it in AMD's overclocking menus.Reply
CPU CO works but if setting both you may need to set it on AMD menus anyway to ensure that it doesn't get reset. -
Kamen Rider Blade This platform wasn't meant to compete with CPU+GPU normal combos.Reply
This APU is optimized for LapTops & tiny NUC's.
That's it, there's only 2x major niches that it works for. -
thestryker The performance with massively tuned DRAM says everything: the 780m is bandwidth starved and they'd likely be better off switching to a quad channel memory architecture. While this is probably relatively viable for laptops I'm not sure it would be on the desktop side of things.Reply