AMD Ryzen Threadripper 2970WX vs Intel Core i5-3470 Comparison

The AMD Ryzen Threadripper 2970WX dominates the Intel Core i5-3470 in every recorded head-to-head benchmark, but the scale of that dominance varies significantly by workload, and the older Intel chip still holds its own in a few narrow, single-threaded contexts relative to its own class. The database records eight head-to-head comparisons, and the Threadripper wins all eight, yet the margin tells a more nuanced story than a simple sweep.

Head-to-Head Benchmarks

The most extreme gaps appear in multi-core rendering tests, where the Threadripper’s core count becomes a decisive factor. In Cinebench R23 multi-core, the AMD part scores 26,305 against the Intel’s 3,957, a delta of -85% from the Intel’s perspective, meaning the Threadripper is roughly 6.6 times faster. The same -85% delta appears across Cinebench R15 multi-core (2,651 vs 398), R20 multi-core (11,048 vs 1,661), and even the single-core variants of those same tests. The consistency of that exact -85% figure across all six Cinebench runs is striking; it suggests that both single-thread and multi-thread performance in this suite scale almost identically between the two architectures, which is unusual.

Single-core performance in Cinebench still favors the Threadripper by that same -85% margin. The i5-3470 scores 56 in R15 single-core and 558 in R23 single-core, while the Threadripper scores 374 and 3,713 respectively. That is a massive per-thread advantage, not just a core-count advantage. The gap narrows considerably in Geekbench, where the Threadripper’s single-core score of 1,236 beats the i5-3470’s 715 by -42.2%, and its multi-core score of 7,195 beats 2,015 by -72%. Those deltas are still large, but they show that the Threadripper’s per-core advantage is smaller in Geekbench than in Cinebench, a useful reminder that benchmark suites measure different facets of a CPU.

A curious detail emerges when looking at the average benchmark scores. The i5-3470 has an average score of 6,906 and sits at the 63rd percentile of all CPUs, while the Threadripper averages 6,760 and sits at the 62nd percentile. Despite losing every head-to-head test, the i5-3470 actually has a slightly higher average score and percentile. The explanation lies in the PassMark tests recorded for the Intel chip, which the AMD part lacks in this dataset. The i5-3470’s PassMark integer math score of 14,840, floating point math of 10,871, and multithread score of 4,669 pull its average up, while the Threadripper only has the eight Cinebench and Geekbench results to contribute. The head-to-head tests are unequivocal, but the overall averages reveal how heavily the choice of benchmarks influences a CPU’s standing.

Architecture Differences

The two processors come from different eras and design philosophies. The i5-3470 is an Ivy Bridge part built on Intel’s 22 nm process, while the Threadripper 2970WX is a Zen+ part built on GlobalFoundries’ 12 nm process. That process difference alone explains much of the efficiency and clock behavior gap, though the Threadripper’s 250 W TDP versus the i5-3470’s 77 W TDP shows the cost of feeding 24 cores.

Core and thread counts are the obvious differentiator: the i5-3470 has 4 cores and 4 threads, while the Threadripper has 24 cores and 48 threads. The Threadripper also carries a much larger cache hierarchy. Its L1 cache is 96 KB per core, its L2 is 512 KB per core, and its L3 is 64 MB shared. The i5-3470 has 64 KB L1 per core, 256 KB L2 per core, and only 6 MB shared L3. The Threadripper’s transistor count is 19,200 million across four 213 mm² dies, while the i5-3470 has 1,400 million on a single 160 mm² die. That is a 13.7 times transistor advantage, though the i5-3470’s smaller die area and older node mean the comparison is not purely about scale.

Memory architecture also diverges sharply. The i5-3470 supports dual-channel DDR3 with a theoretical bandwidth of 25.6 GB/s, while the Threadripper supports quad-channel DDR4 with 93.9 GB/s. Both lack ECC support in this configuration. PCIe lanes differ as well: the Intel part offers Gen 3 with 16 CPU lanes, while the AMD part offers Gen 3 with 60 CPU lanes. The i5-3470 includes Intel HD 2500 integrated graphics, whereas the Threadripper has no integrated graphics at all. The Threadripper’s multiplier is unlocked, allowing overclocking, while the i5-3470’s is locked. Sockets are entirely different: Intel Socket 1155 for the i5-3470, AMD Socket SP3r2 for the Threadripper.

FAQ

Q: Is the AMD Ryzen Threadripper 2970WX faster in every single benchmark?

A: Yes. The database records eight head-to-head benchmarks, and the Threadripper wins all eight. The margins range from -42.2% in Geekbench single-core to -85% in all six Cinebench tests.

Q: Why does the Intel Core i5-3470 have a higher average benchmark score than the Threadripper?

A: The i5-3470 has an average score of 6,906 against the Threadripper’s 6,760. This happens because the database includes many PassMark results for the Intel chip, such as integer math (14,840) and floating point math (10,871), while the Threadripper only has eight Cinebench and Geekbench results. The overall averages are not directly comparable because the test sets differ.

Q: Does the Threadripper win single-core tests by the same margin as multi-core tests?

A: In Cinebench, yes. The delta is -85% for both R15 and R20 and R23 single-core and multi-core. In Geekbench, no. The single-core delta is -42.2%, while multi-core is -72%. The per-thread advantage is smaller in Geekbench.

Q: Can the i5-3470 be overclocked?

A: No. Its multiplier is locked. The Threadripper’s multiplier is unlocked, so it can be overclocked.

Q: Does the i5-3470 have integrated graphics?

A: Yes, it includes Intel HD 2500. The Threadripper has no integrated graphics, so a discrete GPU is required.

Q: Which CPU has more PCIe lanes?

A: The Threadripper has 60 CPU lanes on Gen 3, while the i5-3470 has 16 CPU lanes on Gen 3. That is a 3.75 times difference in lane count.

Specification Differences

The two chips differ in nearly every measurable specification. The i5-3470 has 4 cores and 4 threads, while the Threadripper has 24 cores and 48 threads. Base clocks are 3.20 GHz for the Intel and 3.00 GHz for the AMD, but boost clocks favor the AMD at 4.20 GHz versus 3.60 GHz. TDP is 77 W versus 250 W. The i5-3470 uses Intel Socket 1155, the Threadripper uses AMD Socket SP3r2. Architecture is Ivy Bridge versus Zen+ (Colfax). Process node is 22 nm versus 12 nm. Transistor count is 1,400 million versus 19,200 million, and die size is 160 mm² versus 4x 213 mm².

Cache configurations differ at every level: L1 is 64 KB per core versus 96 KB per core, L2 is 256 KB per core versus 512 KB per core, and L3 is 6 MB shared versus 64 MB. Memory support is DDR3 dual-channel (25.6 GB/s) versus DDR4 quad-channel (93.9 GB/s). PCIe is Gen 3 with 16 lanes versus Gen 3 with 60 lanes. Integrated graphics exist on the Intel part but not the AMD part. The i5-3470 is end-of-life with a launch MSRP of $184, while the Threadripper is active with a launch MSRP of $1299. The i5-3470’s multiplier is locked, the Threadripper’s is unlocked. Release dates are 2012-05-31 for the Intel and 2018-10-01 for the AMD. The part numbers are SR0T8 and YD297XAZUHCAF respectively.

The Verdict

The data points to an unambiguous winner for raw performance: the Threadripper 2970WX is faster in every recorded head-to-head test, and in Cinebench the margin is a uniform -85% across all six variants. Anyone whose workload is dominated by Cinebench-style rendering or multi-threaded Geekbench tasks would see enormous gains from the AMD part. The Threadripper’s 24 cores and 48 threads, 64 MB L3 cache, and quad-channel DDR4 bandwidth at 93.9 GB/s are the structural reasons for that dominance.

However, the i5-3470 is not without a rationale. Its 77 W TDP, integrated graphics, and dual-channel DDR3 setup make it a far lighter system component, and its 63rd percentile standing against all CPUs is actually one point higher than the Threadripper’s 62nd percentile. That percentile gap is a statistical artifact of the different benchmark sets, but it does indicate that the i5-3470 remains competent in the specific PassMark workloads recorded, such as data compression (73,224) and random string sorting (9,004). The Threadripper’s average score of 6,760 places it just below the i5-3470’s 6,906, and its nearest rival, the Intel Core i7-12700E, is only -0.2% away, meaning the Threadripper sits in a very tight competitive cluster. The i5-3470’s nearest rival, the Intel Xeon W-2195, is also 0.2% away, so both chips are near the center of their respective performance neighborhoods.

Where Each One Wins

The Threadripper wins in every scenario that uses the eight head-to-head benchmarks. Cinebench R15, R20, and R23, both single and multi-core, all go to the AMD part by -85%. Geekbench multi-core goes to the AMD by -72%, and Geekbench single-core by -42.2%. For rendering, video encoding, scientific computing, or any heavily threaded workload that scales with cores and memory bandwidth, the Threadripper is the clear choice from the recorded data.

The i5-3470 wins in scenarios not covered by the head-to-head set but present in its individual benchmark list. Its PassMark results show strengths in data compression (73,224), integer math (14,840), floating point math (10,871), and random string sorting (9,004). Those are not compared directly against the Threadripper in this database, so a direct claim of superiority cannot be made, but they do show that the i5-3470 has well-rounded throughput in everyday desktop tasks. The i5-3470 also wins on system-level simplicity: a 77 W TDP, integrated graphics, and a locked multiplier suggest a low-maintenance, low-power build, while the Threadripper’s 250 W TDP and lack of integrated graphics demand a discrete GPU and robust cooling. For a user who prioritizes the specific PassMark workloads and a simpler platform, the i5-3470 remains relevant; for anyone who needs the Threadripper’s multi-core muscle, the head-to-head data leaves no room for doubt.