Which gives you more frame rates when overclocking?

Depends on the load but generally core speed. If at hi-res mem will show significant gains.
 
Hmmm.

I've been trying to find a web site that game bench reviews the core vs mem overclocking of a video card but can't find any.

I game at 4k and my 1080ti is always at 100% when I play Assetto Corsa. I'm getting a decent boost (~73fps jumping to ~78fps) but it doesn't last.

My GPU(MHz) seems to clock down by itself not staying at the high setting?

Code: 
Sample GPU(MHz) Mem(MHz) GPUmV Pwr(%) GPU('C)  GPU%    CPU%     Mem%
2788	2025	11138	1.044	127	69	98	37	39
2789	2025	11138	1.044	130	71	98	47	40
2790	2012	11138	1.044	126	71	97	51	40
2791	2012	11138	1.044	131	72	98	55	40
2792	2012	11138	1.044	124	73	97	57	40
2793	2012	11138	1.044	127	74	97	59	40
2794	2012	11138	1.044	125	75	97	60	40
2795	2012	11138	1.044	127	75	97	61	39
2796	2000	11138	1.031	124	76	97	61	39
2797	2000	11138	1.031	128	76	97	62	39
2798	2000	11138	1.031	123	77	97	63	40
2799	2000	11138	1.031	121	78	97	63	40
2800	2000	11138	1.031	126	78	97	64	40
2801	2000	11138	1.031	126	79	97	63	40
2802	2000	11138	1.031	124	79	97	63	40
2803	2000	11138	1.031	126	80	97	63	39
2804	2000	11138	1.031	125	80	97	63	39
2805	2000	11138	1.031	123	80	97	63	40
2806	2000	11138	1.031	128	81	97	64	40
2807	2000	11138	1.031	126	81	97	64	40
2808	2000	11138	1.031	127	82	98	62	40
2809	2000	11138	1.031	128	82	98	61	40
2810	2000	11138	1.031	123	82	97	60	40
2811	2000	11138	1.031	130	83	97	59	40
2812	2000	11138	1.031	132	83	97	58	40
2813	2000	11138	1.031	130	84	97	59	40
2814	2000	11138	1.031	132	84	97	60	40
2815	2000	11138	1.031	132	84	97	61	40
2816	2000	11138	1.031	132	84	97	63	40
2817	2000	11138	1.031	133	85	97	64	40
2818	2000	11138	1.031	131	85	98	64	40
2819	2000	11138	1.031	133	85	98	63	40
2820	2000	11138	1.031	128	85	97	62	40
2821	2000	11138	1.031	130	86	97	62	40
2822	2000	11138	1.031	130	86	97	62	40
2823	1987	11138	1.025	130	86	97	62	40
2824	1987	11138	1.025	127	86	97	64	40
2825	1987	11138	1.025	132	87	97	64	40
2826	1987	11138	1.025	127	87	97	64	40
2827	1987	11138	1.025	129	87	97	64	40
2828	1987	11138	1.025	130	87	98	64	40
2829	1987	11138	1.025	129	87	97	63	40
2830	1987	11138	1.025	131	87	97	63	40
2831	1987	11138	1.025	127	88	97	63	40
2832	1987	11138	1.025	127	88	98	63	40
2833	1987	11138	1.025	132	88	97	64	40
2834	1987	11138	1.025	131	88	97	64	40
2835	1987	11138	1.025	125	88	98	64	40
2836	1987	11138	1.025	132	88	98	62	40
2837	1987	11138	1.025	131	89	98	62	40
2838	1987	11138	1.025	132	89	97	60	40
2839	1987	11138	1.025	127	89	98	59	40
2840	1987	11138	1.025	129	89	97	59	40
2841	1987	11138	1.025	134	89	97	60	40
2842	1987	11138	1.025	135	89	97	60	40
2843	1987	11138	1.025	132	89	97	59	40
2844	1987	11138	1.025	131	90	97	59	40
2845	1974	11138	1.013	128	90	97	59	40
2846	1974	11138	1.013	134	89	97	59	40
2847	1962	11138	1	129	90	97	60	40
2848	1974	11138	1.013	132	90	97	63	40
2849	1949	11138	0.994	123	89	97	64	40
2850	1987	11138	1.025	129	90	97	64	40
2851	1974	11138	1.013	127	90	98	63	40
2852	1974	11138	1	123	90	97	62	40
2853	1962	11138	0.994	125	90	97	62	40
2854	1962	11138	1	120	89	98	62	40
2855	1962	11138	1	122	89	97	61	40
2856	1936	11138	0.975	122	89	97	62	40
2857	1949	11138	0.994	124	90	97	63	40
2858	1949	11138	0.994	124	89	97	63	40
2859	1949	11138	0.994	123	89	97	64	40
2860	1949	11138	0.994	124	89	97	64	40
2861	1949	11138	0.994	120	90	97	64	40
2862	1949	11138	0.994	120	90	97	64	40
2863	1949	11138	0.994	124	90	97	64	40
2864	1924	11138	0.963	116	89	97	63	40
2865	1936	11138	0.975	118	89	97	64	40
2866	1962	11138	1	126	90	97	64	40
2867	1949	11138	0.994	126	90	98	64	40
2868	1949	11138	0.994	119	90	98	63	40
2869	1949	11138	0.994	126	90	98	63	40
2870	1924	11138	0.963	111	89	97	62	40
2871	1949	11138	0.994	126	90	98	61	40
2872	1924	11138	0.963	115	89	97	60	40
2873	1949	11138	0.994	126	90	97	61	40
2874	1911	11138	0.95	116	89	97	62	40
2875	1936	11138	0.975	124	90	97	63	40
2876	1936	11138	0.975	117	89	97	64	40
2877	1898	11138	0.944	115	89	98	64	40
2878	1924	11138	0.963	117	89	98	64	40
2879	1936	11138	0.975	116	89	97	63	40
2880	1924	11138	0.963	119	89	98	62	40
2881	1924	11138	0.963	113	89	97	63	40
2882	1924	11138	0.963	117	89	98	62	40
2883	1924	11138	0.963	117	89	98	62	40
2884	1924	11138	0.963	117	89	97	64	40
2885	1911	11138	0.95	111	89	97	64	40
2886	1924	11138	0.963	116	89	97	64	40
2887	1949	11138	0.994	122	90	97	65	40
2888	1949	11138	0.994	122	90	98	66	40
2889	1936	11138	0.975	117	89	97	65	40
2890	1911	11138	0.95	111	89	97	64	40
2891	1924	11138	0.963	118	90	98	64	40
2892	1949	11138	0.994	125	90	97	63	40
2893	1949	11138	0.994	126	90	97	64	40
2894	1936	11138	0.975	119	89	98	63	40
2895	1924	11138	0.963	116	89	98	62	40
2896	1924	11138	0.963	112	89	98	61	40
2897	1924	11138	0.963	115	89	98	60	40
2898	1936	11138	0.975	121	90	97	59	40
2899	1924	11138	0.963	116	90	97	59	40
2900	1911	11138	0.95	114	89	98	58	40
2901	1924	11138	0.963	119	89	97	59	40
2902	1911	11138	0.95	113	89	97	60	40
2903	1924	11138	0.963	117	90	97	61	40
2904	1911	11138	0.95	113	89	97	62	40
2905	1924	11138	0.963	116	89	98	62	40
2906	1924	11138	0.963	117	89	97	61	40
2907	1924	11138	0.963	118	89	98	61	40
2908	1924	11138	0.963	117	89	97	61	40
2909	1924	11138	0.963	113	90	98	61	40
2910	1936	11138	0.975	119	90	97	61	40
2911	1936	11138	0.975	119	90	97	63	40
2912	1949	11138	0.994	122	90	95	63	40
2913	1911	11138	0.95	110	89	97	63	40
2914	1924	11138	0.963	116	89	97	64	40
2915	1949	11138	0.994	123	89	97	64	40
2916	1911	11138	0.95	109	89	98	63	40
2917	1924	11138	0.963	118	89	98	61	40
2918	1936	11138	0.975	114	90	97	60	41
 
Newer Nvidia GPUs dynamically reduce core frequency as temperature increases. It looks to me like that's what you're seeing there. There's not much you can do about that besides improve cooling. However, the difference in core frequency you're experiencing is only about 5%, so it shouldn't have a huge impact.

I don't think that there's a lot of testing of core vs memory speeds, because they're not really mutually exclusive. There's nothing stopping you from pushing both of them up and they shouldn't have much effect on the other.
 
Ampere core frequency is also affected by the power envelope. So more power equals more heat so need more cooling to keep the GPU core in a higher boost state.
 
Nagorak said:
I don't think that there's a lot of testing of core vs memory speeds, because they're not really mutually exclusive. There's nothing stopping you from pushing both of them up and they shouldn't have much effect on the other.
Click to expand...

Ah, good to know they don't affect one another.

My case temps are getting pretty high. Since this card is not rear exhausted, it just dumps out hot air inside the case.

Time to revamp my case air flow it seems.
 
Seems like card is ripe for some water cooling action. Throw an AIO on it.
 
Geforce 2s back in the day only benefited from memory OC. Geforce 3 benefited equally from both. Since then, I don't think I've had a card that cared much for memory OC. Even my Fermi 620 (laptop) with a 64bit bus got most of its gains from the core clock.
 
You must log in or register to reply here.
Back
Top