Caching libraries Benchmarks

According to my new library cachebox, I decided to benchmark caching libraries which are I know, to show the power of cachebox ...

If you know other library, tell me to add it to this page.

Note

The cacheing.VTTLCache class was excluded from the benchmark because of numerous unreasonable errors.

🖥️ System Information

• Platform: Linux-6.18.12-1-MANJARO-x86_64-with-glibc2.43
• Python: 3.14.3
• Processor:
• Libraries:
  • cachebox: 6.0.0
  • cachetools: 7.1.4
  • cacheing: 0.1.1
  • lru-dict: 1.3.0

📊 Insertion

• Performance Range: 15.5x difference between fastest and slowest
• Best: dict (72ns, 13.9M ops/sec)

Implementation	Mean (ns)	Std Dev (ns)	Ops/sec	Samples	Stability
dict	72	1	13.9M	97	1.5%
lru.LRU	78	2	12.8M	97	2.2%
cachebox.RRCache	93	2	10.7M	98	2.1%
cachebox.LFUCache	94	2	10.7M	91	1.9%
cachebox.Cache	94	2	10.6M	89	2.1%
cachebox.LRUCache	94	2	10.6M	87	2.0%
cachebox.FIFOCache	95	2	10.5M	99	1.8%
cachebox.TTLCache	162	41	6.2M	907	25.2%
cacheing.RandomCache	266	24	3.8M	859	9.1%
cacheing.LRUCache	294	13	3.4M	863	4.5%
cachetools.Cache	297	13	3.4M	883	4.3%
cacheing.LFUCache	320	21	3.1M	832	6.4%
cachetools.FIFOCache	379	15	2.6M	873	3.9%
cachetools.RRCache	396	27	2.5M	861	6.9%
cachetools.LFUCache	401	17	2.5M	815	4.1%
cachebox.VTTLCache	422	79	2.4M	944	18.7%
moka_py.Moka (lru)	488	145	2.1M	951	29.6%
cachetools.LRUCache	564	21	1.8M	871	3.8%
moka_py.Moka	565	300	1.8M	932	53.1%
cacheing.TTLCache	986	43	1.0M	854	4.4%
cachetools.TTLCache	1,117	56	0.9M	861	5.0%

📊 Lookup Existing

• Performance Range: 13.6x difference between fastest and slowest
• Best: dict (73ns, 13.7M ops/sec)

Implementation	Mean (ns)	Std Dev (ns)	Ops/sec	Samples	Stability
dict	73	1	13.7M	97	1.4%
cachebox.Cache	83	5	12.1M	823	6.2%
cachebox.RRCache	92	1	10.9M	95	1.4%
cachebox.LFUCache	93	1	10.8M	89	1.3%
cachebox.LRUCache	93	1	10.8M	94	1.3%
cachebox.FIFOCache	94	1	10.6M	85	1.2%
cacheing.RandomCache	105	10	9.5M	911	9.9%
lru.LRU	125	8	8.0M	788	6.0%
cachetools.RRCache	154	10	6.5M	894	6.6%
cachetools.Cache	155	11	6.5M	905	7.3%
cachetools.FIFOCache	157	11	6.4M	906	7.0%
cachebox.VTTLCache	173	66	5.8M	956	38.2%
cachebox.TTLCache	174	79	5.8M	985	45.3%
cacheing.LRUCache	191	12	5.2M	806	6.1%
cachetools.LRUCache	306	14	3.3M	851	4.4%
cacheing.LFUCache	315	15	3.2M	866	4.9%
moka_py.Moka (lru)	341	164	2.9M	940	48.0%
moka_py.Moka	349	167	2.9M	951	47.9%
cachetools.LFUCache	392	15	2.6M	877	3.8%
cacheing.TTLCache	681	28	1.5M	893	4.1%
cachetools.TTLCache	994	57	1.0M	951	5.7%

📊 Lookup Missing

• Performance Range: 61.8x difference between fastest and slowest
• Best: dict (76ns, 13.2M ops/sec)

Implementation	Mean (ns)	Std Dev (ns)	Ops/sec	Samples	Stability
dict	76	1	13.2M	85	1.4%
cachetools.FIFOCache	120	8	8.3M	866	6.7%
cachetools.LFUCache	120	8	8.3M	872	7.0%
cachetools.LRUCache	120	8	8.3M	871	7.0%
cachetools.RRCache	122	10	8.2M	897	7.9%
cachetools.Cache	125	12	8.0M	899	9.5%
lru.LRU	188	12	5.3M	871	6.5%
moka_py.Moka	332	149	3.0M	940	44.7%
moka_py.Moka (lru)	341	156	2.9M	940	45.8%
cacheing.RandomCache	467	16	2.1M	861	3.4%
cacheing.LFUCache	574	18	1.7M	811	3.1%
cachetools.TTLCache	653	22	1.5M	887	3.4%
cacheing.LRUCache	715	20	1.4M	882	2.8%
cachebox.Cache	1,012	394	1.0M	867	39.0%
cacheing.TTLCache	1,697	37	0.6M	918	2.2%
cachebox.RRCache	3,842	96	0.3M	882	2.5%
cachebox.VTTLCache	4,146	107	0.2M	913	2.6%
cachebox.LRUCache	4,166	133	0.2M	892	3.2%
cachebox.LFUCache	4,206	143	0.2M	943	3.4%
cachebox.TTLCache	4,639	111	0.2M	979	2.4%
cachebox.FIFOCache	4,683	125	0.2M	926	2.7%

📊 Deletion

• Performance Range: 8.6x difference between fastest and slowest
• Best: dict (63ns, 15.9M ops/sec)

Implementation	Mean (ns)	Std Dev (ns)	Ops/sec	Samples	Stability
dict	63	4	15.9M	822	6.6%
cachebox.Cache	71	2	14.1M	865	3.2%
cachebox.RRCache	72	3	13.8M	819	4.2%
cachebox.FIFOCache	82	3	12.1M	848	3.5%
lru.LRU	88	4	11.3M	829	4.8%
cachebox.LRUCache	99	8	10.1M	817	8.4%
cachebox.LFUCache	101	8	9.9M	804	8.1%
cachebox.TTLCache	108	5	9.3M	886	4.3%
cachebox.VTTLCache	128	9	7.8M	816	7.3%
cachetools.Cache	162	10	6.2M	930	6.0%
cacheing.LRUCache	199	12	5.0M	926	5.8%
cachetools.FIFOCache	236	12	4.2M	900	5.0%
cachetools.LRUCache	238	15	4.2M	948	6.1%
cacheing.LFUCache	259	15	3.9M	933	5.8%
cachetools.LFUCache	270	16	3.7M	853	6.1%
cacheing.RandomCache	288	34	3.5M	913	11.9%
cachetools.RRCache	318	39	3.1M	937	12.3%
moka_py.Moka	467	82	2.1M	889	17.5%
moka_py.Moka (lru)	468	76	2.1M	878	16.2%
cachetools.TTLCache	510	18	2.0M	877	3.5%
cacheing.TTLCache	539	26	1.9M	878	4.7%

📊 Popitem

• Performance Range: 3972.9x difference between fastest and slowest
• Best: cachebox.TTLCache (893ns, 1.1M ops/sec)

Implementation	Mean (ns)	Std Dev (ns)	Ops/sec	Samples	Stability
cachebox.TTLCache	893	204	1.1M	973	22.8%
cachebox.FIFOCache	918	187	1.1M	970	20.4%
lru.LRU	989	129	1.0M	930	13.0%
cachebox.LRUCache	1,073	111	0.9M	881	10.4%
cacheing.LRUCache	1,387	178	0.7M	955	12.8%
cachetools.FIFOCache	1,906	246	0.5M	951	12.9%
cacheing.LFUCache	1,984	300	0.5M	971	15.1%
cachetools.LRUCache	2,090	287	0.5M	979	13.7%
cacheing.TTLCache	2,095	299	0.5M	952	14.3%
cacheing.RandomCache	2,799	403	0.4M	963	14.4%
cachetools.RRCache	3,052	383	0.3M	968	12.5%
cachebox.RRCache	7,212	3,757	0.1M	998	52.1%
cachetools.LFUCache	16,119	242	0.1M	894	1.5%
cachebox.LFUCache	31,176	848	0.0M	911	2.7%
cachebox.VTTLCache	37,553	958	0.0M	922	2.6%
cachetools.TTLCache	3,548,048	162,835	0.0M	747	4.6%

📊 Hash Collision

• Performance Range: 7.3x difference between fastest and slowest
• Best: cachebox.Cache (3838ns, 0.3M ops/sec)

Implementation	Mean (ns)	Std Dev (ns)	Ops/sec	Samples	Stability
cachebox.Cache	3,838	2,129	0.3M	1000	55.5%
cachebox.RRCache	3,894	2,142	0.3M	994	55.0%
cachebox.VTTLCache	4,181	2,296	0.2M	989	54.9%
cachebox.LRUCache	4,206	2,350	0.2M	980	55.9%
cachebox.LFUCache	4,279	2,362	0.2M	990	55.2%
lru.LRU	4,434	2,446	0.2M	960	55.2%
cacheing.RandomCache	4,494	2,451	0.2M	999	54.5%
dict	4,514	2,466	0.2M	1000	54.6%
cachebox.FIFOCache	4,537	2,519	0.2M	1000	55.5%
cachebox.TTLCache	4,625	2,626	0.2M	989	56.8%
moka_py.Moka (lru)	6,532	3,512	0.2M	999	53.8%
moka_py.Moka	6,588	3,495	0.2M	986	53.1%
cachetools.FIFOCache	8,858	4,887	0.1M	999	55.2%
cachetools.RRCache	8,911	4,946	0.1M	1000	55.5%
cachetools.Cache	8,935	4,965	0.1M	999	55.6%
cacheing.LRUCache	8,936	4,924	0.1M	1000	55.1%
cacheing.TTLCache	9,413	4,898	0.1M	984	52.0%
cachetools.LRUCache	17,836	9,988	0.1M	997	56.0%
cachetools.TTLCache	18,490	9,945	0.1M	993	53.8%
cacheing.LFUCache	23,930	13,365	0.0M	997	55.8%
cachetools.LFUCache	28,031	15,270	0.0M	998	54.5%

🚀 Usage

Prerequisites

uv sync

Run Benchmarks

# All benchmarks
python main.py
# Specific benchmark
python main.py insertion
python main.py lookup_existing

Understanding Results

• 🟢 Excellent (A+): <2% noise ratio - Highly reliable measurements
• 🟡 Good (A-B): 2-10% noise ratio - Acceptable reliability
• 🔴 Poor (C-D): >10% noise ratio - Less reliable measurements

Performance Metrics:

• Mean (ns): Average execution time in nanoseconds
• Ops/sec: Operations per second (throughput)
• Stability: Measurement consistency (lower noise = more stable)